12e

MICROWAVE CIRCUITS
K. C. GUPTA
HAKI CEBI
CHOONSIK CHO
ZHIPING FENG
University of Colorado at
Boulder
The term microwave circuits is used to identify the elec-
trical circuits used at microwave frequencies for perform-
ing signal processing functions like amplication,
frequency conversion, mixing, detection, phase shifting,
ltering, and power dividing. By microwave frequencies,
we refer to electromagnetic signals whose wavelength is in
centimeters, roughly from 30 cm to 1 mm with the corre-
sponding frequencies ranging from 1 GHz (GHz 10
9
Hz)
to 300GHz. The frequency range from 30 to 300 GHz is
also known as the millimeter-wave band. Microwave fre-
quencies present several interesting and unusual features
not found in other portions of the electromagnetic fre-
quency spectrum. These features make microwaves
uniquely suitable for several useful applications in tele-
communications, radar, industrial heating and sensors,
and so on. The most common consumer application of mi-
crowaves is the domestic microwave oven used for food
processing.
Microwave circuits differ from lower-frequency elec-
tronic circuits for several reasons. Active devices (transis-
tors, diodes, etc.) used at microwave frequencies are
special designs and in several cases operate on entirely
different physical principles. Parasitic reactances associ-
ated with passive and active circuit elements used at low-
er frequencies become significant and can cause
disastrous effects on performance of circuits at microwave
frequencies. Dimensions of lumped elements used in low-
frequency electronics can become comparable to the wave-
lengths at microwave frequencies and cause what are
known as distributed circuit effects. Transmission lines
(and other structures) used for transmission of signals
from one location to another inside a circuit need to be
designed differently from those at lower frequencies. All
these features make the design, technology, and operation
of microwave circuits significantly different from their
lower-frequency counterparts.
1. TYPES OF MICROWAVE CIRCUITS
Types of microwave circuits may be classied in several
different ways. Depending on the type of active devices
used, there are vacuum-tube circuits and solid-state cir-
cuits. Depending on the technology used, there are printed
circuits, hybrid integrated circuits, monolithic circuits,
and so on. Depending on the circuit functions, there are
amplier circuits, lter circuits, mixer circuits, power di-
vider circuits, and so on. Depending on the special perfor-
mance features, there are low-noise circuits, high-power
circuits, and so on. Depending on the transmission struc-
tures whose sections form the basic building blocks for
circuit design, there are waveguide circuits, coaxial cir-
cuits, stripline circuits, microstrip circuits, coplanar wave-
guide (CPW) circuits, nline circuits, slotline circuits,
dielectric waveguide circuits, and so on. Depending on
the special material properties used, there are ferrite cir-
cuits, surface acoustic wave (SAW) circuits, and so on.
This article provides a brief overview of the important
types of microwave circuits. Several specific types of mi-
crowave circuits are described in more detail in other ar-
ticles in this encyclopedia. These include articles on
frequency converters/mixers, microwave signal ampliers,
microwave couplers, microwave detectors, microwave l-
ters and multiplexers, microwave oscillators, microwave
phase shifters, microwave switches, power combiners/
dividers, and stripline components.
2. LUMPED AND DISTRIBUTED CIRCUITS
All electronic circuits can be grouped into ve classes de-
pending on their physical dimensions compared to the
wavelength at the frequency of operation. When all three
physical dimensions of a component or a circuit are much
smaller than the wavelength at the frequency of opera-
tion, we call it a lumped circuit. These are most exten-
sively used components and circuits at lower frequencies.
Lumped components are used at microwave frequencies
also, but their dimensions have to be proportionately
smaller. When one of the physical dimensions of a compo-
nent is comparable to the wavelength (other two being still
small), we refer to these as one-dimensional components,
and the circuits using these components are called
one-dimensional circuits. Circuits using sections of trans-
mission lines as components, commonly known as
transmission-line circuits, are one-dimensional. Transmis-
sion-line circuits are used extensively at microwave and
millimeter-wave frequencies. Components and circuits
with two of their dimensions comparable to the wave-
length at the operating frequency are appropriately called
two-dimensional components. Thin-lm components,
planar components and circuits fabricated on thin sub-
strates, microstrip patch antennas, and reduced-height
waveguide components belong to this class. These planar
components are the important building blocks in micro-
wave integrated circuits. The fourth class of electronic cir-
cuits have all three of their dimensions comparable to the
wavelength at the operating frequency. These are known
as three-dimensional components. Waveguides, hollow
metallic cylindrical tubes used as transmission structures
in place of conventional transmission lines, have both of
their transverse dimensions comparable to the wave-
length. Circuits using these waveguide sections and reso-
nant cavities made out of these waveguides are examples
of three-dimensional components and circuits used at mi-
crowave frequencies. Finally, the fth class of electronic
circuits (classied by their size) have at least one of their
dimensions much larger than the wavelength at the oper-
ating frequency. These are known as quasioptical circuits.
The term quasioptical is derived from their likeness to
optical circuits that are orders of magnitude larger than
the submicron wavelengths at optical frequencies.
2704 MICROWAVE CIRCUITS
Previous Page
3. EVOLUTION OF MICROWAVE CIRCUITS
For a long time the term microwave circuits was synon-
ymous with waveguide circuits. A waveguide was recog-
nized as a useful transmission structure for microwave
frequencies in the early 1930s. The work of Southworth
and others [13] at Bell Telephone Laboratories deserves
mention in this respect. It was soon realized that a short
length of waveguide, with suitable modications, might
function as a radiator and also as a reactive element. Res-
onant cavities and horn antennas are mentioned in a 1936
article by Southworth [1]. Modern waveguide circuitry
had its beginning in the efforts to obtain both a more ef-
cient transfer of microwave power from a source to a
waveguide transmission line, thereby providing the ele-
ments of a transmitter, and again in the efcient recovery
of microwave power at the receiving end, thereby provid-
ing the elements of a receiver. These efforts led to the de-
velopment of several components like traveling detectors,
wavemeters, terminations, and so on. Some idea of the
techniques used in 1934 can be obtained by recalling that
optical benches were commonly used to set up microwave
experiments [4]. Several photographs of equipment of
those days are available in an interesting article survey-
ing the history of the progress of microwave arts published
in the ftieth anniversary issue of Proceedings of the
IRE [4].
The principle of multiple reections from discontinu-
ities and the associated principle of cavity resonance
played an important role in the development of micro-
wave technology. In some cases, these principles were
used to match a source of microwave power to a wave-
guide. In others, they served to match a waveguide to a
receiver, such as a crystal detector. In still others, they
served to pass freely a band of frequencies. Together, these
principles formed the foundations of microwave circuits.
One of the key features of microwave circuits has been the
empirical adjustment or tuning of characteristics by
screws and irises (and even by denting) in waveguides.
In the beginning it was an art that was learned by trial
and error. This came to be known as plumbing and had
been for quite a long time a practical tool for microwave
engineers.
Perhaps the greatest single contribution to the engi-
neering analysis of microwave circuits was by Phillip H.
Smith [5], who provided a graphical tool for solving oth-
erwise complicated transmission-line problems. Not only
were laborious calculations avoided, but, while solving the
problems on a Smith chart, one could visualize the step-
by-step processes underway. Few gadgets of microwave
circuitry have been more useful than the Smith chart.
Rapid developments in microwave circuits took place dur-
ing World War II, when special laboratories were set up at
the Massachusetts Institute of Technology and at Colum-
bia University to apply microwave techniques to radar
problems. Many significant developments in microwave
circuits took place during these years, but were published
later. A few of those deserve mention. Fox [6] developed
devices by which phase could be added progressively to a
waveguide. Another product was a hybrid tee (or magic
tee) [7] and still another equally significant one was the
rst directional coupler [8]. All these devices found prac-
tical uses immediately. Another direction of wartime evo-
lution was the extension of lter techniques to higher
frequencies, leading to transmission line lters. Simulta-
neously, analytical tools were also developed. The classical
description of network performance in terms of voltages,
currents, impedance, and admittance matrices was re-
placed by a description based on the transmitted and the
reected wave variables, leading to the concept of scatter-
ing matrix.
The scattering matrix formalism allows simpler repre-
sentation of multiport microwave networks. At this stage
in the development of microwave circuits, two basic trans-
mission structures were employed frequently. These were
the waveguide and the coaxial TEM-mode line. Wave-
guides provided higher power capability and low loss
that lead to high-Q resonant cavities. Coaxial lines pro-
vided inherently wider bandwidth because of the absence
of dispersion effects. Also, the concept of impedance could
be easily interpreted in the case of coaxial lines. This sim-
plied the design of components. These two transmission
structures (waveguides and coaxial lines) grew as impor-
tant components for microwave circuits. Often their roles
have been complementary; sometimes they appear in the
same module. It was at this stage that a very useful mi-
crowave technique emerged from a special adaptation of
two-conductor transmission-line theory. Introduced by
Barrett and Barnes [9] in 1951, this structure, as used
presently, consists of a thin strip of conductor sandwiched
between two dielectric plates metalized on the outside.
This structure is known as a stripline. Early stripline
work used razor blades and glue to cut the thin strips and
paste them on dielectric sheets. With the availability of
copper-clad laminates (rst introduced for printed cir-
cuits) the stripline techniques have developed into a pre-
dictable and precise batch process technology. The rst
detailed account of stripline circuits was made available
by the Sanders Tri-plate manual [10], published in 1956. A
comprehensive account of stripline circuits was made
available in a book by Harlan Howe, Jr. [11]. The most
significant feature of a stripline transmission structure is
that the characteristic impedance of the line is controlled
by the width of the central strip, which is fabricated by
photoetching a copper-clad dielectric substrate. The two-
dimensional nature of the stripline circuit conguration
permits the interconnection of many components without
the need to break the outer conductor shielding. This also
allows the placement of the input and output ports with a
high degree of exibility. Striplines were found to be very
convenient for use in parallel-line couplers because of the
natural coupling between two strips placed close to each
other. The principles of the coupled-line directional cou-
pler were introduced by Wheeler [12] in 1952. Even today
a vast majority of directional couplers use a stripline con-
guration.
In the early 1950s, another type of transmission struc-
ture was conceived [13,14], consisting of a single dielectric
laminate with a conducting strip on one side and a com-
plete conducting coating on the other side. This structure
is known as a microstrip line. Microstrip lines enjoyed a
brief spell of popularity and intensive investigations in the
MICROWAVE CIRCUITS 2705
1950s, but were not readily accepted at that time for mi-
crowave use because of the high loss per unit length
caused by radiation. This was largely a result of the low
dielectric constant (about 2.5) of the substrate materials
then in use. Further developments were prevented by the
lack of availability of both the high-dielectric-constant,
low-loss materials and suitable methods for processing
and production.
Ever-increasing demands for miniaturized microwave
circuitry for use in weapons, aerospace, and satellite ap-
plications led to renewed intense interest in microstrip
circuits in the 1960s. An elegant analysis of microstrip
structure based on conformal mapping transformation
was presented by Wheeler [15,16]. The technology of
high-dielectric-constant, low-loss dielectric materials and
that of deposition of metallic lms were perfected [17] and
became easily available in the late 1960s. This led to rapid
developments in the use of microstrip lines in microwave
circuits. Today, microstrip line is the most common trans-
mission structure used in hybrid and monolithic micro-
wave integrated circuits.
The availability of a planar microwave transmission-
line structure like microstrip line, coupled with the rapid
developments in microwave semiconductor devices and
the techniques of thin-lm deposition and photolithogra-
phy, eventually resulted in the technology of microwave
integrated circuits [1821]. Microwave integrated circuits
(MICs) represent an extension of thin-lm hybrid inte-
grated circuit technology to microwave frequencies. These
hybrid MICs consist mostly of passive components and
circuits in the form of conducting patterns deposited on
ceramic or dielectric substrates plus active devices mount-
ed on these circuits in the form of chips or in specially de-
signed packages. In addition to microstrip, other types of
lines called slotline and coplanar lines [22,23] have been
used in some MICs. Slotline consists of a slot in the con-
ducting pattern on one side of a dielectric substrate. The
other side of the substrate does not contain any metal-
lization. Coplanar lines also involve a metallization pat-
tern, but only on one side of the substrate.
Another trend in microwave circuits is the use of
lumped elements. Previously, lumped elements could not
be used because the size of available lumped elements was
comparable to the wavelength at microwave frequencies.
With the use of photolithography and thin-lm tech-
niques, the size of elements (capacitors, inductors, etc.)
can be reduced so much that these elements can be used
up to J-band (1020GHz) frequencies [24,25]. Use of
lumped elements on dielectric substrates, along with
semiconductor devices in chip form mounted thereon, is
an attractive option for microwave integrated circuits.
Cost reduction of the order of one-ftieth or more has
been predicted with the use of these types of circuits [24].
Apart fromreduction in size, there is another advantage of
lumped elementscircuit design and optimization tech-
niques perfected at lower frequencies can now be directly
used in the microwave frequency range. In addition to
lumped elements and one-dimensional transmission-line
components, two-dimensional planar components have
also been proposed for use in microwave circuits [26].
These components are compatible with stripline and
microstrip line and provide a useful alternative in micro-
wave circuit design.
The current generation of MICs is monolithic micro-
wave integrated circuits (MMICs) using semiconductor
substrates [27,28]. Semiconductor substrates used are
high-resistivity gallium arsenide and, to a limited extent,
high-resistivity silicon. Difculties arise from the need to
use a variety of microwave semiconductor devices that
cannot be fabricated by a common process, as well as be-
cause of the requirement of large substrate areas when
distributed elements (transmission-line sections) are used
for passive functions. GaAs technology [29] and GaAs met-
al semiconductor eld-effect transistors (MESFETs) [30]
play the key role in microwave monolithic integrated
circuits.
Microwave integrated circuits (hybrid or monolithic)
exhibit almost the same advantages as those available in
the case of integrated circuits at lower frequencies [31],
namely: (1) improved system reliability, (2) reduced vol-
ume and weight, (3) batch production, and (4) eventual
cost reduction when a large number of standardized items
are required.
As in the case of low-frequency integrated circuits, the
MICs are responsible for both the expansion of present
markets and the opening of many new applications, in-
cluding a host of nonmilitary uses.
There are some difculties associated with the use
MICs [31]. Before MICs became popular, the microwave
circuit designers and users had the exibility to incorpo-
rate tuners and adjustment screws in circuits in order to
optimize the performance of the circuit after fabrication.
MICs, especially if they have to meet high reliability stan-
dards, lack these trimming arrangements. Consequently,
devices used in MICs need to be characterized precisely
and the circuits have to be designed more accurately.
Computer-aided design, simulation, and optimization
[32,33] techniques have therefore become a necessity.
The current state of the art in microwave circuits is
summarized in several recent books [3436].
4. WAVEGUIDE CIRCUITS
As pointed out in Section 3, hollow metallic single-con-
ductor waveguides of rectangular and circular cross sec-
tion were among the earliest forms of transmission
structures used at microwave frequencies (see Fig. 1). A
rectangular waveguide has been more popular and is still
used today for many applications. A large variety of wave-
guide circuit components such as couplers, detectors, iso-
lators, attenuators, and slotted lines are commercially
available for various standard waveguide frequency bands
ranging from 1 GHz to over 220GHz [37,38,42]. Because of
the more recent trend toward miniaturization and inte-
gration, more and more microwave circuits are currently
fabricated using planar transmission lines (such as mi-
crostrip line and coplanar waveguides) discussed later in
this article. However, there is still a need for waveguide
circuits in many applications such as high-power systems,
millimeter-wave systems, and some precision test/mea-
surement applications [38,42]. Also waveguides can be
combined with other kinds of transmission lines [39] for
some special applications.
4.1. Modes in a Waveguide
Unlike most of the other transmission structures dis-
cussed later in this article, waveguides do not support
the transverse electromagnetic (TEM) mode of wave prop-
agation. Waveguides support two other kinds of modes
known as the transverse electric (TE) and transverse mag-
netic (TM) modes. TE modes have their electric eld com-
ponents only in a plane transverse to the direction of
propagation along the waveguide, and for TM modes the
magnetic eld components are totally in the transverse
plane. Both of these types of modes have cutoff frequencies
below which wave propagation is not possible. For a rect-
angular waveguide the mode with the lowest cutoff fre-
quency is the TE
10
mode. Field patterns of the TE
10
mode
for a rectangular waveguide are shown in Fig. 2. Various
eld components for this mode can be expressed as
[34, pp. 145146]
E
y
E
y0
sinpx=ae
jbz
H
x
H
x0
sinpx=ae
jbz
H
z
H
z0
cospx=ae
jbz
1
where a is the waveguide width (in the x direction) and b is
the phase constant of the wave along the z direction given
by
b
k
2
p
a
_ _
2
_
2
where k is the wavenumber ko

me
p
. The cutoff fre-
quency for the dominant TE
10
mode is given by
f
c10

1
2a

me
p 3
4.1.1. Dispersion Characteristics. The fact that the
propagation constant for individual waveguide modes is
a nonlinear function of frequency, and that the different
modes start to propagate at different frequencies, leads to
wave dispersion in a waveguide [38, pp. 106107].
4.1.2. Power-Handling Capability. The power-handling
capability for a transmission medium needs to be charac-
terized for high-power microwave circuits. Metallic wave-
guides can handle very high power because of their
physical structure. In a rectangular waveguide operating
in the fundamental mode, the maximum peak power that
the waveguide can handle is given by [40]
P
max
416ab kW=cm
2
4
where b is the waveguide dimension in the y direction.
4.2. Waveguide Circuit Components
Circuits for all kinds of signal processing functions have
been designed using waveguides. A few of these are re-
viewed here.
4.2.1. Waveguide Phase Shifters. A phase shifter is a
circuit that produces an adjustable shift in the phase an-
gle of the wave transmitted through it. There are different
types of waveguide phase shifters. Two of these, linear and
rotary phase shifters, are described here:
Linear Phase Shifter. An example of linear phase shift-
ers is the circuit consisting of three dielectric slabs
placed in a rectangular waveguide [41] as shown in
Fig. 3. The center slab is free to move longitudinally,
and it is moved by a suitable drive mechanism to
x
a
z
y
z
0
(a)
(b)

r
b
a
[
Figure 1. Hollow metallic single-conductor waveguides: (a) rect-
angular waveguide; (b) circular waveguide. Waveguides were the
earliest form of transmission structures used at microwave
frequencies and are still used today for special applications.
b b
x
x
z
y
y
y
E
H
z
a
g/2
g/2
Figure 2. Field patterns of TE

10
mode in a rectangular wave-
guide. Dashed lines show the H eld, and solid lines show the E
eld. Note that there is no z component of electric eld.
which it is keyed by means of a dielectric key that
protrudes through a long centered slot cut in one
broad face of the guide. Each end of the dielectric slab
is cut stepwise to provide a broadband multisection
quarter-wave transformer to match the partially
lled guide to the empty and completely lled guide.
If the center slab is displaced a distance x to the right,
the effect is to lengthen lines 1 and 3 by an amount x
and to shorten lines 2 and 4 by the same amount x.
Therefore, the phase shifter change undergone by a
wave propagating through the structure is
Df b
1
b
3
b
2
b
4
x 5
The phase shift Df is proportional to the displace-
ment x. The amount of phase shift increases if the
dielectric constant of the slab is increased. If a mate-
rial whose e
r
equal to 2.56 is used in a 3-cm wave-
guide of dimensions a2.25cm, the phase shift
obtained is about 0.4 rad/cm of displacement. About
16 cm of displacement gives a phase shift of more
than 3601.
Rotary Phase Shifter. The rotary phase shifter [42, pp.
262266] is a better precision instrument than the
linear phase shifter. It consists of a half-wave plate
and two quarter-wave plates (see Fig. 4). The quarter-
wave plate on the left converts a linearly polarized
TE
11
mode into a circularly polarized mode, and the
quarter-wave plate on the right produces a linearly
polarized wave when a circularly polarized wave is
incident on it. Rotation of the half-wave plate through
an angle y changes the phase of the transmitted wave
by an amount 2y. This simple dependence of the
phase change on mechanical rotation is the main fea-
ture of the rotary phase shifter.
4.2.2. Microwave Hybrid Junction. A rectangular
waveguide hybrid, which is more popularly known as
magic tee, is shown in Fig. 5. If a wave in the dominant
TE
10
mode is incident at the port 4, the structure is sym-
metrical with respect to this wave, and hence equal pow-
ers are transmitted to port 1 and 3. If E
n
m
represents the
transmitted electric eld in the nth port when the incident
wave is in the mth port, then E
1
4
E
3
4
. Besides, it can be
seen that no power is transmitted to port 2 from port 4,
that is, E
2
1
0. On the other hand, if a TE
10
wave is inci-
dent at port 2, the E eld has an odd symmetry about the
plane of symmetry and therefore excites elds in ports
1 and 3 which are 1801 out of phase. Hence, E
1
2
E
3
2
.
Also, power incident at port 2 is not transmitted to port 4.
Therefore, E
4
2
0. The power coupling factor may not be
exactly one-half if ports 2 and 4 have reections. To ensure
a coupling of exactly one-half, which is desirable, the junc-
tion needs to be matched by irises or probes.
Dielectric key
Line 1, [
1
Line 2, [
2
Line 3, [
3
Line 4, [
4
Dielectric slabs
(a)
(b)
0.3a a
Figure 3. A linear phase shifter in a waveguide conguration.
The dielectric key is used to move the central dielectric slab and
thereby change the phase shift. (From Ref. 42, r Ellis Horwood
Limited, reprinted with permission.)
x
y
Dielectric slab
Rotating section
(a)
(b)
/2 plate /4 plate /4 plate z z z
([
1
[
2
)l = / 2
Figure 4. (a) A rotary phase shifter consists of a l/2 plate and
two l/4 plates (rotation of the l/2 plate changes the phase; l/4
plates convert circular polarization into linear and vice versa); (b)
details of the l/4 plate. (From Ref. 53, r McGraw-Hill, 1992, re-
printed with permission.)
2
4
3
Symmetry plane
1
Figure 5. A microwave hybrid or magic tee in a rectangular
waveguide conguration. The magic tee is a directional coupler
with 3 dB coupling and is commonly used in balanced mixers.
4.2.3. Directional Couplers. A directional coupler is a
four-port circuit. An example is shown in Fig. 6. A portion
of the wave incident at port 1 couples into the bent wave-
guide through the hole a
1
. The remaining wave travels to
the hole a
2
in the main waveguide (ports 1, 2), and a por-
tion of it again couples into the bent waveguide. If the
magnitudes of the energy coupled through holes a
1
and a
2
are equal, and if the distance between a
1
and a
2
is l
g
/4,
then the two coupled signals are reinforced at port 4 be-
cause they arrive with equal phase. On the other hand,
the two coupled signals cancel each other at port 3, be-
cause they are 1801 out of phase. Similarly, a wave
incident on the junction from port 2 travels to ports 1
and 3, but no signal travels to port 4. The ratios of signal
ow between ports 1 and 4 and between ports 2 and 3 are
known as the coupling coefcient of the directional cou-
pler. In general, the leakage of energy through holes a
1
and a
2
is kept quite small. A directional coupler can be
used as a standing-wave detector and forms an important
component in microwave and millimeter-wave network
analyzers. Some other designs [37, Sect. 7.2, 42, pp.
267271] of typical waveguide directional couplers are
shown in Fig. 7.
4.2.4. Waveguide Filters. Design procedures for lters
using waveguides and other transmission structures have
certain common features. These lters can be designed
using the low-frequency prototype lter synthesis tech-
niques. One such technique, called the insertion-loss
method, begins with a complete specication of the atten-
uation characteristics of the lter and develops a basic
prototype lowpass lter having the desired passband char-
acteristics. Using suitable frequency transformations and
element realizations, the required type of lter (lowpass,
bandpass, highpass, or bandstop lters) is derived. The
lumped element values of these lters are then realized in
terms of the distributed circuit elements. Special features
for implementing lters in waveguide conguration are
reviewed in this section:
Waveguide Stub Filters. One of the simplest realiza-
tions of waveguide bandpass (or bandstop) lters is a
waveguide stub lter [38, pp. 185190] shown in
Fig. 8. For bandpass lters, the short-circuited
E-plane stubs are a half-wavelength long at the cen-
ter frequency of the lter, and are connected in cas-
cade with l/4 separations. Band rejection lters use
stubs that are an odd multiple of quarter-wavelength,
also separated by l/4 distance. The stubs can be
asymmetric or symmetric along the main waveguide
and with or without steps. The selection of the stub
conguration depends on the required bandwidth
and power-handling capability. Filters with stepped
stubs can be designed to yield narrower bandwidth.
Filters with stubs on both sides of the waveguide
broadwall have large power-handling capability.
E-Plane Filters. E-plane lters have been developed as
compatible ltering structures for integrated milli-
meter-wave circuits [43,44] and are most often real-
ized in nline techniques or as all-metal structures.
Their common feature is that the lter metallization
pattern is obtained using photolithographic tech-
niques. Thus, the geometry of the component is real-
ized with very small manufacturing errors, which is
very important at millimeter-wave frequencies. Four
different congurations of E-plane waveguide lters
are shown in Fig. 9. All of types are designed as band-
pass lters [43,45,46]. In Fig. 9, the shaded part is
dielectric substrate to support the thin metallization
structures; t
m
is the thickness of the metal layer and
t
d
is the thickness of the dielectric substrate; and l
c
is
the length of the metal strip and l
r
is the gap between
metal strips.
4

[
a
2
a
1
3
[ + /2
[ + / 2
z
g
/4
2 1
Figure 6. A waveguide directional coupler is a widely used cir-
cuit component used to sample the waves traveling in one par-
ticular direction (say, from 1 to 2) independent of the reected
wave traveling in the opposite direction (from 2 to 1).
Rotated auxiliary guide
Main guide
Main guide
Coupling holes
g
/4
Auxiliary
guide
(a)
(b)
Coupling hole
To receiver
To absorber
Figure 7. Two other designs of waveguide directional coupler

circuits. In both of these cases, coupling takes place through holes
in the common wall of the two waveguides.
Corrugated Waveguide Filters. Corrugated waveguide
structures similar to that shown in Fig. 10 are used as
lowpass lters in numerous antennas feed system to
reject the spurious harmonics from transmitters [38,
pp. 200207; 49]. They can also be designed as band-
pass lters, with a wide or narrow passband response
[47,48].
Evanescent-Mode Waveguide Filters. Evanescent-mode
wave-guide lters (Fig. 11) can be designed [5052] to
provide a very wide stopband with low passband
insertion loss. Their size is compact, even when the
passband is located in the lower microwave frequency
region. For these reasons the evanescent-mode wave-
guide structures are often used as pseudolowpass
lters, or as bandpass lters in a wide microwave
spectrum. Because of the high skirt selectivity
achievable in this lter type, they can also be used
in diplexers and multiplexers.
4.3. Multiplexers
Muliplexer circuits are required for combination or sepa-
ration of communication channels at different frequencies.
The multiplexer for the antenna feed systems must pro-
vide separation of the receive and transmit bands and
combination of the individual transmission channels that
cover only a small portion of the frequency band. There
are four different multiplexing methods that are applied in
feed systems [38, pp. 252307]: (1) the circulator/lter
chain, (2) the directional lter approach, (3) the manifold
multiplexing techniques, and (4) the branching lter con-
cept. Each of these has its own particular properties and
applications.
4.3.1. Waveguide Circuits Using Active Devices. Wave-
guide circuits using active devices can be designed for
various applications, such as oscillator, mixer, detector,
and so on [37, pp. 325490]. A waveguide cavity Gunn os-
cillator [53] is shown in Fig. 12 as an example. In this de-
sign the high impedance of the waveguide is transformed
into low impedance at the location of the Gunn device by
means of quarter-wave transformers. The cavity resonant
frequency can be adjusted by changing the location of the
b
b
b
b
(a)
w
k
(b)
2
(c)
b
b
2
z
0
z
0
z
0
z
0
4
2
Figure 8. Waveguide stub lters: (a) asymmetric stubs without
steps; (b) asymmetrical stubs with steps; (c) symmetric stubs
without steps. Both bandpass and bandstop characteristics can be
designed in these circuits.
(a) (b)
(c) (d)
a a
t
d
l
r
l
c
l
r
l
c
t
m
a
t
m
a
b
Figure 9. E-plane bandpass lters: (a) large gap nline lter;
(b) single metal insert lter; (c) double metal insert lter; (d) tri-
ple metal insert lter. In these designs, n dimensions and hence
RF performance can be controlled accurately by photolithography
techniques. (From Ref. 38, rArtech House, 1993, reprinted with
permission.)
b
Figure 10. Longitudinal cross section of a corrugated waveguide
lter.
b
a
(a)
(b)
Figure 11. Ridged waveguide evanescent mode bandpass lter:
(a) top view; (b) side view. These designs provide compact lters at
lower microwave frequencies.
short circuits. A tuning screw can be used for ne tuning
of the cavity. More examples of active wave guide circuits
can be found in Refs. 37 and 53.
4.4. Computer-Aided Design of Waveguide Circuits
Waveguide circuits require an approach for computer-aid-
ed design (CAD) different from that used for transmission-
line circuits at microwave frequencies. The traditional
CAD methods for waveguide circuit are usually based on
the network analysis, with various discontinuities in the
waveguide modeled separately by a combination of equiv-
alent reactances [54]. This equivalent circuit approach has
some drawbacks. The models are valid only for a specied
geometry and only within a certain range of parameters.
Another problem associated with using the equivalent cir-
cuit models is their inability to account for higher-order
mode-coupling effects, which can occur if discontinuities
are in close proximity. A eld-theory-based approach [55]
overcomes these limitations. Some features of this ap-
proach are a very accurate prediction of frequency re-
sponses, higher-order mode effects taken into account, no
restrictions on the wavelength (or frequency range), and
straightforward extension in millimeter-wave bands. Sev-
eral numerical methods are used for eld analysis of wave-
guide circuits; the most popular are the nite-different
time-domain method (FDTD) [56] nite-element method
(FEM) [57], transmission-line matrix method (TLM) [58],
mode-matching techniques (MMT) [59], and the method of
integral equations [60]. For automated design and yield
analysis of waveguide circuits, modal analysis [61] has
emerged as the most useful electromagnetic simulator,
either in the generalized scattering matrix (GSM) formu-
lation or in the generalized admittance matrix (GAM)
form. It has been demonstrated [61] that for waveguide
circuits the GAM approach requires only half the number
of unknowns at the internal ports and hence is much more
efcient than the GSM representation.
5. COAXIAL-LINE CIRCUITS
Coaxial line is the most commonly used transmission
structure over a very wide range of frequencies from
very low frequencies through microwave frequencies and
extending into the millimeter-wave frequency range.
However, because of the convenience of physical size, co-
axial line circuits are popular only in the microwave fre-
quency range. They are too bulky at lower frequencies and
very difcult to fabricate (as well as very lossy) at milli-
meter-wave frequencies.
The geometry of a coaxial line is shown in Fig. 13. One
can derive the expressions for electromagnetic elds in
this line solving Laplace equation for scalar potential. As
shown in Fig. 13, the inner conductor is considered to be at
V
0
volt potential and the outer conductor is at 0 V. The
electric and magnetic eld vectors, can be derived as [34]
Er; f; z
V
0
^ rre
jbz
rlnb=a
6
Hr; f; z
V
0
^
ffe
jbz
Zr lnb=a
7
where b o

me
p
and Z

me
p
are the phase constant and
the intrinsic impedance of the medium, respectively.
Coaxial lines possess general properties of TEM-mode
transmission lines. Characteristic impedance Z
0
of a co-
axial line lled with a dielectric material of relative
dielectric constant e
r
as shown in Fig. 13, is
Z
0

60
e
r
p ln
b
a
O 8
In addition to TEM modes, coaxial lines can also support
TE and TM waveguide modes. When coaxial line dimen-
sions are selected appropriately for the operating frequen-
cy range; these modes are evanescent modes and they are
excited only near discontinuities or sources. In practice, it
is essential to know the cutoff frequencies of the lowest-
order waveguide mode and use the coaxial line below this
frequency. Various types of microwave circuits can be re-
alized using coaxial lines [62,63]. However, with the more
recent advances in planar circuit technology and because
of their size and fabrication difculties, they are not used
commonly. In the past, coaxial lines have been widely used
to design passive lter circuits. A common type of high-
pass lters constructed by coaxial lines as shown in Fig. 14
Output
waveguide
Two-section
quarter-wave
transformer
Gunn
device
Tuning screw
V
Figure 12. A Gunn oscillator waveguide circuit that uses two
quarter-wave sections to transform the high impedance of the
waveguide to a low impedance at the Gunn device. (From Ref. 53,
r McGraw-Hill, 1992, reprinted with permission.)
V= 0
y
V = V
0
x
z
a
b

j
[
Figure 13. Geometry of a coaxial line. The ratio of the outer to
inner conductor radii (b/a) determines the characteristic imped-
ance Z
0
of the line.
has been described in the classic book by Matthaei et al.
[62]. In this conguration, coaxial stubs present shunt in-
ductances, and disks spacers constitute series capacitors.
Another coaxial lter example described by Matthaei
[62] is a series capacitance-coupled half-wave resonator
circuit shown in Fig. 15. This lter is realized by breaking
the inner conductor at several locations. The gap spacing
needed to produce a desired coupling can be found either
experimentally or theoretically.
Coaxial-line components are also used extensively as
coaxial probes connectors in between various circuit as-
semblies and for connecting circuits to instrumentation,
and so on. Most of the coaxial lines that are used as cables
and connectors have a 50 O characteristic impedance ex-
cept for 75 O coaxial cable used for television systems. Co-
axial connectors must have low standing-wave ratio
(SWR), no spurious higher-order modes, mechanical
strength, and repeated usability. Some of the most com-
mon microwave coaxial connectors are popularly known
as type N connector (originally named after P. Neill),
(SMA) SubMiniature Amphenol connector, (SSMA) Scaled
SubMiniature Amphenol connector, and (APC-7) Amphe-
nol Precision Connector, 7-mm connector. These connector
types are shown in Fig. 16. The type N connector is a rel-
atively large connector with an outer diameter of 0.625 in.
The recommended upper operating frequency ranges from
11 to 18 GHz. The SWR is typically less than 1.07. The
SMA connector is small compared to the type N connector
with an outer diameter of the female end of 0.210 in. and
can be used up to 25 GHz. SMA connectors modied to
work up to 40 GHz are known as K connectors. An SSMA
connector is even smaller. The outer diameter of the fe-
male end is about 0.156 in. and the maximum operating
frequency is about 38 GHz. The APC-7 connector is a pre-
cision connector which has an SWR less than 1.04 and an
operating range of up to 18 GHz. Coaxial connectors are
described in Refs. 34 (pp. 169170) and 64.
6. STRIPLINE CIRCUITS
6.1. Striplines
A stripline [11,65] is a planar-type transmission line that
lends itself well to microwave integrated circuitry and
package feedthroughs. The geometry of a stripline is
shown in Fig. 17a. A thin conducting strip of width W is
centered between two wide conducting ground planes with
a separation b. The entire region between the ground
planes is lled with a dielectric.
Unlike microstrip lines and other open planar trans-
mission lines described later in this article, a stripline can
support a pure TEM mode because it has a homogeneous
dielectric medium. The stripline, however, can also sup-
port higher-order TM and TE modes. These modes can be
suppressed with shorting screws between the two ground
planes and by restricting the ground planes spacing to less
than one quarter wavelength. A sketch of the eld lines for
the TEM stripline mode is shown in Fig. 17b.
A B
A
Section B-B Section A-A
B
C
Figure 14. A highpass lter constructed by coaxial lines.
Lumped series capacitors and shunt inductors provide the lter
operation. (From Ref. 62, r Artech House, 1980, reprinted with
permission.)
Figure 15. Series capacitance coupled half-wave resonators l-
ter. Series coupling gaps are located in between cascaded straight
resonator elements. The lter is realized by breaking the inner
conductor at several locations.
a
b
c
(a)
(b)
(c)
Figure 16. Some of the most common microwave coaxial connec-
tors: (a) type Nconnector; (b) SMA connector; (c) APC-7 connector.
6.2. Stripline Parameters
An exact solution of Laplace equation of electromagnetic
elds in a stripline can be obtained by the conformal map-
ping approach [66]. However, closed-form expressions that
give a good approximation of the exact results are used in
circuit design [34].
The phase constant for a stripline is given (by the usual
relation for homogeneously lled lines) as
b
o
v
p
o

m
0
e
0
e
r
p

e
r
p
k
0
9
where o is the angular frequency, v
p
is the velocity of the
wave along the line, m
0
is the permeability of free space, e
0
is the permittivity of free space, e
r
is the dielectric constant
of the stripline dielectric, and k
0
is the phase constant of
free space.
The characteristic impedance of a transmission line is
given by
Z
0

L
C
_

1
v
p
C
10
where L and C are inductance and capacitance per unit
length of the line. An approximate expression for charac-
teristic impedance [34] of striplines is
Z
0

30p
e
r
p
b
W
e
0:441b
11
where
W
e
b

W
b

0 for W=b > 0:35
0:35 W=b
2
for W=bo0:35
_
Since the stripline is a TEM-mode line, the attenuation
due to the dielectric loss is obtained by the procedure com-
monly used for other TEM lines [67]. The attenuation due
to the conductor loss is approximated as
a
c

2:710
3
R
s
e
r
Z
0
30pb t
A for

e
r
p
Z
0
o120 Np=m
0:16R
s
Z
0
b
B for

e
r
p
Z
0
> 120 Np=m
_
_
12
with
A1
2W
b t

1
p
b t
b t
ln
2b t
t
_ _
B1
b
0:5W0:7t
0:5
0:414t
W

1
2p
ln
4pW
t
_ _
where t is the thickness of the strip metallization.
6.3. Examples of Stripline Circuits
Design procedures for stripline circuits are identical to
those for other TEM mode transmission line circuits. The
main difculty in transferring the design from one kind of
transmission line (say, coaxial line) to another (say, strip-
line) arises from the fact that discontinuity and junction
reactances are different for different kinds of transmission
structures. Quite often, the rst-order designs are carried
out without considering the effect of discontinuity/junction
reactances. Then the circuit performance is computed tak-
ing discontinuity/junction reactances into account, and
designable parameters of the circuit are optimized to com-
pensate for discontinuity/junction effects. This design
methodology is common to microwave circuit design us-
ing any kind of transmission structure.
6.3.1. Branchline Directional Couplers. These couplers,
similar to the one shown in Fig. 18 [68], are essentially
power division networks with two important features
namely, the two ports are mutually isolated (ports 1 and 4
in Fig. 18 when the input signal is connected to port 1),
and the output signals at the other two ports (ports 2 and 3
in Fig. 18) are out of phase by 901. These circuits form
building blocks of several other circuits such as balanced
mixers, variable attenuators, pin-diode phase shifters, di-
rectional lters, diplexers, multiplexers, and transmitre-
ceive (TR) switches. Branchline couplers are also used
Ground
planes
E
H
c
r
(a)
(b)
z
x
y
W
b
Figure 17. (a) Geometry of striplinea thin strip of width W is
inserted in a dielectric with ground planes on the top and the
bottom; (b) electric and magnetic elds in a stripline. Striplines
support a pure TEM mode.
4
P P
1
3
Y
c
V
a
V
b
V
b
V
a
Y
c
Y
c
Y
c
2
= 90

= 90
Figure 18. A branchline coupler using striplines. Port 1 is the
input port, ports 2 and 3 are output ports, and port 4 is the iso-
lated port. Output signals at ports 2 and 3 are 901 out of phase.
extensively in antenna array feed networks in preference
to Y-junction-type power dividers and as impedance trans-
formers. In active circuits, they provide the advantage of
direct-current (DC) coupling for biasing. Branchline cou-
plers are forward-wave couplers and hence can be cascad-
ed without crossing lines.
6.3.2. Parallel-Coupled-Line Directional Couplers. Paral-
lel-coupled-line directional couplers shown in Fig. 19 [69]
offer much larger bandwidths as compared with the
branchline couplers. They are mostly backward-wave cou-
plers, although forward-wave couplers are also possible
using an inhomogeneous medium. The most commonly
used parallel-coupled directional coupler is the TEM-mode
single-section backward-wave coupler. As the term back-
ward-wave coupler implies, the electric and magnetic
eld interaction between the parallel-coupled conductors
causes the coupled signal to travel in a direction opposite
to that of the input signal. Maximum coupling occurs
when the length of the coupling region is equal to one-
quarter wavelength (or an odd multiple of quarter wave-
length). Analysis of the couplers is carried out in terms of
two normal modes of propagation known as even and odd
modes for symmetrical couplers. Even and odd modes ex-
hibit even and odd symmetry of elds with respect to the
plane of symmetry. These couplers offer a perfect match
and innite directivity at all frequencies because of the
inherent property that the even- and odd-mode phase ve-
locities are equal when the propagating mode is a pure
TEM.
6.3.3. Hybrid Rings. The branchline coupler as well as
the coupled-line backward-wave coupler provides a phase
difference of 901 between the two outputs. For hybrid ring
couplers shown in Fig. 20 [70], the two output signals are
either in-phase or 1801 out of phase depending on the
choice of the input port. The circumference of the ring is
3l/2, where l is the wavelength in the stripline at mid-
band frequency. For an input at port 1, outputs at ports
2 and 4 are 1801 out of phase and the port 3 is isolated
(with no output). When the input is at port 2, the two out-
puts at ports 1 and 3 are in phase and the port 4 is isolated.
6.3.4. Power Dividers. In several microwave applica-
tions (as, e.g., a feed for a phased-array antenna) the in-
put signal is required to be divided into several
equiamplitude, equiphase output signals. A matched,
symmetric n-way power divider has the advantage that
it gives neither amplitude nor phase imbalance at any
frequency. Such a power divider can also be used as an
n-way power combiner by simply reversing the input and
output ports. Using such combiners, output powers of a
number of solid-state ampliers and oscillators can be
combined over a wide frequency range. A two-way power
divider stripline circuit is shown in Fig. 21 [71]. The re-
sistor R between the output port ensures input match at
ports 2 and 3 when the circuit is used as a power combiner.
6.3.5. Filters. Stripline lters generally make use of a
cascade of distributed circuit elements in the form of cou-
pled resonators, stubs, and so on. As pointed out in the
discussion for waveguide lters, stripline lters are also
designed using low-frequency prototype lter synthesis
techniques. A coupled-line bandpass lter is shown in
Fig. 22 [72]. This conguration is also commonly used
for other lters using planar lines such as microstrip lines
and coplanar waveguides.
6.3.6. Pin-Diode Switches. A pin diode (which consists
of an intrinsic layer sandwiched between p- and n-type
1
2
Y
c
Y
c
Y
c
Y
a
Y
a
Y
b
Y
b
Y
c
P
P
4

3
0
/4 z
0
/4
z
z
0
/4 z
0
/4
3
Figure 20. A stripline ring hybrid. The circumference is 3/2
wavelengths. When a signal is fed to port 1, output signals at
ports 2 and 4 are 1801 out of phase, and port 3 is isolated. For an
input signal fed to port 3, output signals at ports 2 and 4 are in
phase and port 1 is isolated.
Coupled
power P
3
= P
1
P
2
Isolated
port P
4
= 0
1
3
2
4
w
w
1
w
(l =
0
/4)
s
Input power
P
2
P
1
Output power
Point of symmetry
s
0
c
0
c
r
w w
Figure 19. A parallel-coupled-line coupler real-
ized in stripline structure. Equal widths are used
for two strips. When a signal enters into port 1,
port 2 is the direct output and port 3 is the cou-
pled point. No signal comes out of port 4, which is
called the isolated port.
layers) acts as an electronic switch when operated at the
forward and reverse bias states [73]. As a basic switching
element, it is extensively used in the realization of multi-
ple-throw switches, phase shifters, modulators, limiters,
and duplexers. A single-pole single-throw switch using a
pin diode is shown in Fig. 23 [65]. With a variable forward
bias, the forward bias resistance of the pin diode can be
varied over a wide range. This property is used in realizing
electronically variable attenuators. Pin-diode circuits can
also be used as attenuators by varying the forward bias
current of the diodes. Pin-diode attenuators with constant
input impedance characteristic can be built by incorporat-
ing a circulator or a hybrid coupler in the circuit. Figure 24
shows a hybrid-coupled pin-diode attenuator [65].
6.3.7. Phase Shifters. Phase shifters can be built using
pin diodes, varactors, or GaAs eld-effect transistors
(FETs). Of these three semiconductor devices, pin diodes
are the most commonly used because of reproducibility of
their characteristics and high power-handling capability.
Digital phase shifters are extensively employed in phased
arrays to electronically scan the radiated beam. These
phase shifters can be broadly classied as either the re-
ection type or the transmission type. The reection-type
phase shifters can be realized using a circulator or a
hybrid coupler. Some common designs for transmission-
type phase shifters are the switched line, the loaded line,
and the lowpass highpass. For phased array applications,
several of these circuits are cascaded to form multibit
phase shifters.
Microwave phase shifters can be designed using
various different kinds of transmission lines. Figure 25
shows a switched-line phase shifter with series-mounted
diodes [74].
6.3.8. Mixers. A mixer circuit is an essential compo-
nent of almost all receivers used in communication, radar,
and radioastronomy applications. Microwave mixers make
use of nonlinear semiconductor devices, usually Schottky
barrier diodes for mixing operation. A typical mixer con-
sists of a nonlinear mixer diode together with coupling
networks for feeding the signal [radiofrequency (RF)] and
local-oscillator (LO) power and for extracting the (IF)
signal. Practical mixer congurations can be broadly di-
vided into three categories: single-ended mixers, balanced
1
2 Z
c
Z
c
Z
c
Z
c

3
2
R = 2Z
c
2 Z
c
z
0
/4
z
0
/4
Figure 21. A matched stripline two-way power divider. Input
power at port 1 splits equally into output ports 2 and 3. A resistor
(R2Z
c
) ensures that the circuit is matched at ports 2 and 3.
Z
c
S
n 1, n
s
01
s
12
S
n,

n + 1
Z
c
l
l
l
l
Figure 22. A stripline parallel-coupled band-
pass lter. Each coupled section comprises one
quarter-wavelength coupled lines (l l/4). Z
c
is
the characteristic impedance of input/output
ports.
Z
c
Z
l
Z
p
V
C
b
C
b

z
0
/4
z
0
/4
z
0
/4

Z
Z
c
Figure 23. Layout of series SPST (single-pole single-throw)
switch with the biasing circuit. dc blocking capacitors and dc re-
turn transmission inductor are used, and RF bypass transmission
capacitor of a quarter-wavelength has low characteristic imped-
ance (Z
p
o25O). (From Ref. 65, r New Age Int. Ltd., 1989, re-
printed with permission.)
mixers, and double-balanced mixers. Of the various types
of mixers, the balanced mixer employing Schottky barrier
diodes is the most commonly used conguration in prac-
tical application. Figure 26 shows a hybrid ring balanced
mixer [65].
Several other types of stripline circuits have been re-
ported in literature [11,34,65]. Microwave circuits de-
scribed in this section can also be realized in microstrip
conguration, described in the next section.
7. MICROSTRIP CIRCUITS
Microstrip line is the most frequently used planar trans-
mission structure and forms the basic building block for
almost all hybrid microwave integrated circuits (MICs)
and monolithic microwave integrated circuits (MMICs).
The physical geometry of a microstrip line is shown in
Fig. 27a, and the approximate eld distribution is depicted
in Fig. 27b. Microstrip line consists of a single dielectric
substrate with a complete conducting coating (ground
plane) on one side and a conductor strip on the other side
of the substrate. In contrast to a stripline, the top surface
of the microstrip circuitry is open. Because of the nonho-
mogeneous dielectric medium surrounding the conductor
strip (the substrate and the air above), the microstrip line
cannot support a pure TEM mode. However, a quasi-TEM
mode approximation is used for analysis of microstrip lines
and is adequate for design of microstrip circuits [23].
Popularity of microstrip circuits is due to an increasing
trend in miniaturization of, and cost considerations for,
microwave circuits. Compared to waveguides, and coaxial
lines, and striplines, it is easier to fabricate microstrip
circuits and integrate them with other active and passive
microwave devices. The open nature of the microstrip
lines allows easy access to circuitry to mount active de-
vices and lumped element components such as resistors,
Out
Z
c
Z
c
Z
c
Z
c
Z
c
Z
c
In
Z
c
/ 2

z
0
/4
Z
c
/
Z
c
Z
c
Z
Z
4
3
1
2
2
0
0
z
0
/4
z
0
/4
Figure 24. A hybrid-coupled pin-diode attenuator. The attenua-
tion is controlled by varying the forward bias current of the di-
odes. Two identical diodes are mounted symmetrically in the two
output arms of the 3 dB, 901 hybrid coupler and are shunted by
matched loads. (From Ref. 65, r New Age Int. Ltd., 1989,
reprinted with permission.)
l
0
Z
c
Z
c
Z
c
D
3
D
4
1 2 D
1
D
2
Z
c
l
1
Figure 25. A switched-line phase shifter with series mounted
diode to switch between two xed transmission-line paths. The
differential phase shift is b(l
1
l
0
).
Ground
IF
3
4
3 dB, 180 hybrid
1
2
D
1
T
1
F
1
T
2
F
2
+
D
2
+
LPF
RF
LO
Figure 26. A hybrid ring balanced mixer where D
1
,D
2
are mixer
diodes; F
1
,F
2
are image reject lters; and T
1
,T
2
are matching
transformers. The RF power fed to port 1 and the LO power fed to
port 3 are split equally between the output ports 2 and 4. At the
two mixer diodes, the RF signals appears 1801 out of phase with
each other whereas the LO signals appear in phase. (From Ref.
65, r New Age Int. Ltd., 1989, reprinted with permission.)
x
E
H
W
r
h
(a)
(b)
y
c
Figure 27. Geometry of a microstrip line: (a) geometry (h is sub-
strate height, W is the width of the conducting strip, and e
r
is
relative dielectric constant); (b) electric and magnetic eld distri-
bution in a microstrip-line quasi-TEM mode.
capacitors and inductors. However, microstrip line circuits
have some disadvantages such as higher loss, radiation,
dispersion, and spurious coupling among components
when compared to coaxial-line circuits.
7.1. Design Formulas for Microstrip Lines
Analysis and design considerations for microstrip lines
are very well documented in the literature [16,23,75,76].
7.1.1. Effective Dielectric Constant e
eff
. Because of the
nonhomogeneous dielectric structure of microstrip lines,
the concept of effective dielectric constant has been intro-
duced [16] and is commonly used for calculating line wave-
length and phase velocity as needed in microstrip circuit
design. A formula that is commonly used for calculation of
e
eff
is [75]
e
eff

e
r
1
2

e
r
1
2
110h=W
1=2
13
where h is height of the dielectric substrate, Wis the width
of the conducting strip, and e
r
is the dielectric constant of
the substrate.
7.1.2. Characteristic Impedance (Z
0
). As for any other
transmission line, the characteristic impedance for the
quasi-TEM mode is the most important parameter in mi-
crostrip circuit design. Given the dimensions h and Wand
the value of e
r
, the characteristic impedance of a micro-
strip line can be found by using the following formulas [35,
pp. 5253]:
For narrow strips (W/ho3.3):
Z
0

119:9
2e
r
1
_ ln 4
h
W

16h=W
2
2
_
_ _ _
1
2
e
r
1
e
r
1
_ _
ln
p
2

1
e
r
ln
4
p
_ __
14
For wide strips (W/h43.3):
Z
0

119:9p
2

e
r
p
W
2h

ln 4
p

lnep
2
=16
2p
e
r
1
e
2
r
_ _ _
e
r
1
2pe
r
ln
pe
2
ln
W
2h
0:94
_ _ _ __
1
15
where e is the exponential base; e 2.71828. For
given values of characteristic impedance and e
r
, the
W/h ratio can be found by using the following ex-
pressions.
For narrow strips (when Z
0
> 44 2e
r
O):
W
h

e
A
8

1
4e
A
_ _
1
16
where
A
Z
0
2e
r
1
_
119:9

1
2
e
r
1
e
r
1
_ _
ln
p
2

1
e
r
ln
4
p
_ _
For wide strips (when Z
0
> 44 2e
r
O):
W
h

2
p
_ _
fB 1 ln2B 1g
e
r
1
pe
r
lnB 1 0:293
0:517
e
r
_ _
17
where B59:95p
2
=Z
0
e
r
p
.
7.2. Microstrip Discontinuities
Microstrip circuits, like other types of microwave circuits,
contain transmission line discontinuities such as bends,
width changes, gaps, and junctions. These discontinuities
introduce parasitic reactances and cause a degradation in
circuit performance. Various types of microstrip disconti-
nuities which are encountered in microwave circuits are
shown in Fig. 28. This gure also includes approximate
lumped-element models for these discontinuities. At high-
er microwave frequencies, these discontinuity reactances
become significant and need to be taken into account.
Closed-form relations for discontinuity model elements
(a)
Z
0
Z
01
Z
02
C
P
C
P
C
G
Z
0
C
Z
0
Z
01
Z
01
Z
01
L
1
L
2
L
3
C
Z
02
Z
02
Z
03
Z
03
Z
01
Z
02
C
Z
02
L L
Z
0
(b)
(c)
(d)
Figure 28. Some common types of microstrip discontinuities and
their equivalent lumped-element models: (a) open-ended micro-
strip; (b) gap in microstrip; (c) change in width. (d) T junction.
and their values form a key part of any microwave circuit
computer-aided design (CAD) software. However, equiva-
lent lumped circuit models are not available for all types of
discontinuities and substrate types. Another difculty is
that the accuracy for discontinuity models degrades at
higher frequencies. Lack of accurate discontinuity models
at millimeter-wave frequencies is still the main bottleneck
in the CAD for millimeter-wave circuits [77].
In order to compensate for the discontinuity effects, one
can construct the equivalent circuit for the discontinuity
and take it into account in the design process. The second
approach is to minimize the discontinuity effect by mod-
ifying the geometry of the discontinuity, such as chamfer-
ing or mitering the strip conductor in case of microstrip
right-angle bends. Compensation techniques for disconti-
nuities have been reported in the literature [35, pp. 140
165].
7.3. Passive Microstrip Circuits
Some commonly used microstrip passive circuits are cou-
plers, power dividers, impedance matching circuits, and
lters. Design process for these circuits is similar to that
for stripline circuits.
7.3.1. Couplers. As in case of stripline circuits, the two
common types of microstrip line couplers are also coupled
line directional couplers and branchline directional cou-
plers. Layouts of these couplers are similar to those of
stripline couplers. Even and odd mode analysis technique
is also applied in the design of microstrip couplers. For
coupled line couplers in microstrip conguration, the
phase velocities for even and odd modes are not equal.
This factor limits the directivity of these couplers. Design
of these couplers is well explained in Refs. 34 and 35 .
Branch line couplers are also similar to corresponding
circuits in stripline congurations. They can be designed
for different values of the coupling factor and may have
more than two branches [34,35] in order to increase the
circuit bandwidth.
7.3.2. Filters. Various types of lters [34,35] can be re-
alized by using microstrip lines. Again, their design meth-
odology is similar to those of stripline lters. Lowpass
lters can be formed with cascaded sections of microstrip
lines. One rst designs the prototype lter using lumped
elements and then substitutes these elements with their
microstrip-line equivalents. A short (ol
g
/4) length of a
high-impedance line behaves like a series inductance. Also
a very short 5l
g
=4 length of low-impedance line behaves
like a shunt capacitance. An example of a lowpass lter
constructed by microstrip line elements is shown in
Fig. 29.
Microstrip bandpass and bandstop lters can also be
realized by using coupled sections of microstrip lines. Two
popular types of microstrip bandpass lters are end-cou-
pled and parallel (edge)-coupled bandpass lters. Figure
30 shows a general layout for an end-coupled microstrip
bandpass lter. In this circuit, coupling gaps are located in
between and couple the cascaded microstrip resonator
elements. The gap width is usually much smaller than
the substrate height to ensure the required coupling be-
tween the two adjacent resonators. Parallel-coupled band-
pass microstrip lters are similar to the stripline lter
shown in Fig. 22. Bandpass and bandstop lters [34,35]
can also be realized using quarter-wave open-circuited or
short-circuited microstrip line resonators.
7.4. Active Microstrip Circuits
Microstrip lines have been used in practically all possible
types of active circuits used at microwave frequencies.
7.4.1. Microstrip Ampliers. Microstrip lines are exten-
sively used in various types of microwave amplier cir-
cuits [34,78,79]. All microwave amplier circuits require
some type of matching circuits both at the input and out-
put sides in order to obtain specied gain characteristics
over a desired frequency range. Microstrip lines are used
for realizing matching networks as well as for biasing net-
works for microwave ampliers. Microstrip lines are also
used in distributed ampliers with multioctave band-
widths [78,79]. Various types of microstrip matching net-
works are: single-stub matching network, double-stub
matching network, quarter-wave transformer, multisec-
tion transformer, and tapered line. Microstrip amplier
circuit design examples including matching and biasing
network design considerations are available in Refs. 34,
78, and 79.
7.4.2. Microstrip Oscillators. There are numerous types
of oscillator circuits. One of the most common types of
L
2
L
4
L
6
Z
0
Z
l
Z
h
Z
l
Z
h
Z
l
Z
h
Z
0
C
1
C
3
C
5
l
1
l
2
l
3
l
4
l
5
l
6
(a)
(b)
(c)
Figure 29. Filter design example: (a) lowpass lter prototype
circuit using lumped elements; (b) stepped-impedance implemen-
tation; (c) microstrip layout of nal lter; high-impedance lines
(narrow lines) behave like inductances, and low-impedance lines
(wide lines) behave like capacitances.
Figure 30. General microstrip layout for an end-coupled band-
pass lter (series coupling gaps are located in between cascaded
straight resonator elements).
microstrip oscillator circuits is the dielectric resonator
oscillator (DRO). Layout of a microstrip DRO oscillator
[35,80] is shown in Fig. 31. In this circuit, the dielectric
resonator is coupled to two microstrip lines used to pro-
vide a feedback path from the drain to the gate of the
MESFET. A microstrip single-stub matching circuit is
used for the output matching.
7.4.3. Active Microwave Filters. Active microwave tun-
able lters have been reported in the literature [35,81].
Figure 32 shows a layout of a varactor-tuned, multipole
active microwave bandpass lter described in Ref. 81.
7.4.4. Microstrip Circuits for High-Speed Digital Cir-
cuits. Microstrip lines are also used in ECL high-speed
circuits or GaAs integrated circuits (ICs) for interconnec-
tion among the components and/or device chips. These
microstrip lines are formed by the conductors of integrat-
ed circuits printed on a circuit board. At high frequencies
or high speeds, it is essential to minimize reection at the
interconnections. Also, crosstalk between two parallel in-
terconnects can become significant. Various design tech-
niques for microstrip interconnections in high-speed
digital circuits are described in the literature [35,82].
7.5. Monolithic Microstrip Circuits
Monolithic microwave integrated circuits (MMICs) are
microwave circuits in which all circuit components (active
and passive) are fabricated on the same semiconductor
substrate [83]. MMIC circuits are used in many areas of
microwave circuits with an increasing popularity because
of their significant advantages over hybrid MICs in terms
of lower cost, smaller size, better performance, and higher
reliability. In the lower microwave frequency range,
lumped elements are used for realizing microwave match-
ing networks. However, in the higher frequency range
(over 20 GHz), lumped elements become lossy and difcult
to design, and distributed elements such as microstrip and
coplanar waveguides are used.
Monolithic microwave circuits are described in a sepa-
rate article in this encyclopedia.
8. COPLANAR WAVEGUIDE CIRCUITS
8.1. Coplanar Waveguides
Unlike microstrip lines, the ground plane in a coplanar
waveguide (CPW) [8486] is located on the same side of
the substrate that contains the strip conductor. CPWs, co-
planar strips, and slotlines are categorized as coplanar
lines or uniplanar lines [85] because all the metallization
is contained in a single layer. CPW congurations have
been widely utilized to realize a variety of microwave cir-
cuits including capacitors, inductors, magic tees, mixers,
lters, oscillators, resonators, distributed ampliers, and
so on. Congurations of coplanar strips, CPW, and slot-
lines are shown in Fig. 33.
8.1.1. Field Distribution in Coplanar Lines. Knowledge of
eld distribution in transmission lines is useful to micro-
wave circuit designers because it helps in conguring lo-
cation and orientation of lumped active and passive
elements in transmission line circuits. Approximate elec-
tric eld and magnetic eld distributions in three coplanar
transmission structures are shown in Fig. 34. The eld
distributions are different from those in a microstrip line,
and they lead to some advantages these lines exhibit com-
pared to microstrip lines.
8.2. Advantages of CPW Circuits
One of the advantages of CPW circuits over microstrip
circuits arises from the fact that mounting of lumped com-
ponents in shunt connection is easier in a CPW. In this
case, drilling of holes through the substrate is not needed
to reach the ground plane. Also, transition from a CPW to
a slotline is easier to fabricate. This allows a great exi-
bility in the use of mixed transmission media.

1

4

3

5

6
DR
2
Figure 31. Microstrip implementation of the parallel feedback
DRO. A dielectric resonator is coupled to two microstrip lines used
to provide a feedback path from the drain to the gate of a
MESFET [80].
L
o
w
-
p
a
s
s

f
i
l
t
e
r
L
o
w
-
p
a
s
s

f
i
l
t
e
r
Varactor
Varactor
. . .
L
s
L
s
Figure 32. The multipole active tunable lter. The negative re-
sistance is realized by the MESFET circuit. (From Ref. 81, r
IEEE, 1990, reprinted with permission.)
CPW circuit congurations can be designed to exhibit a
lower sensitivity to substrate thickness, less dispersion
effect, and lower losses than the corresponding microstrip
circuits. Implementation of circuits in CPW congurations
allows thick substrates to be used, thus avoiding the need
to use fragile thin substrates at higher microwave and
millimeter-wave frequencies, as occurs for microstrip cir-
cuits. Since active components can easily be inserted in
CPW circuits also, CPW congurations are used increas-
ingly in monolithic microwave and millimeter-wave inte-
grated circuits. As strip width and gap dimensions in CPW
can be reduced without changing the CPW characteristic
impedance, CPW circuits can be designed to radiate much
less energy than the corresponding microstrip circuits. In
addition, dispersive effects in the CPW can be reduced by
choosing smaller transverse dimensions.
The CPW congurations, however, have some disad-
vantages such as possible excitation of the slot mode, low-
er power-handling capability, and eld nonconnement.
Slot mode is an alternative mode possible in a CPW when
two ground planes are not at the same potential and E
elds in two slots are oriented in the same direction. This
mode can be excited by a nonsymmetric excitation (such as
caused by a tee-junction discontinuity). Airbridges are
needed to suppress excitation of the slot mode. The elds
in a CPW are less conned than those in microstrip lines,
and therefore they make the CPW circuits more sensitive
to packaging covers or shields placed above the circuit.
8.3. CPW Circuit Design Considerations
The main difference in the design of CPW circuits as com-
pared to microstrip circuits is the different characteristics
of discontinuities occurring in CPW circuits. As the den-
sity of active and passive devices in CPW circuits increas-
es, the population of discontinuities also increases. Thus,
the slot mode can be excited more frequently because of
asymmetric structure. When MMICs are composed of
more than one kind of transmission lines, there is the
need for appropriate transitions such as CPW to slotline or
CPW to microstrip line. In the design of CPW circuits,
characterization of these discontinuities and transitions
needs to be taken into account for obtaining the desired
performance. For the analysis and modeling of CPW dis-
continuities, several numerical methods like mode-match-
ing method [87], nite-difference method [88], spectral
domain analysis [89], transmission-line method [90],
x
x
y
z
(a)
(b)
(c)
z
y
w w s
t
w w s
h
h
b
h w
r
c
r
c
Figure 33. Various types of coplanar lines. (a) Coplanar strip.
(b) Coplanar waveguide. (c) Slotline.
Magnetic field lines
Electric field lines
(a)
(b)
(c)
c
r
c
r
k
x
x
x
Figure 34. Electrical and magnetic eld distributions in (a) co-
planar strips, (b) a coplanar waveguide, and (c) a slotline.
integral equation method [91], and method of lines [92]
have been used.
In addition to discontinuities, other factors such as dis-
persion, metallization thickness, dielectric loss, conductor
loss, radiation, and surface wave loss also affect the per-
formance of CPW circuits. All these factors need to be
taken into account in the design of CPW circuits.
8.4. Examples of CPW Circuits
Most of microwave circuits developed using microstrip
lines and other transmission lines can be realized in
CPW congurations also. Some examples of CPW circuits
are reviewed in this section.
8.4.1. Capacitors and Inductors. CPW circuits need the
basic reactive elements, namely capacitors and inductors,
to be realized in CPW congurations. Due to the exibility
provided by CPW congurations to accommodate lumped
elements both in series and the shunt connections, capac-
itors and inductors may be implemented in many ways.
Series or shunt circuits for capacitors or inductors are
possible using CPWor CPW-slotline transitions, and some
of these are shown in Fig. 35 [23].
8.4.2. Filters. Parallel-coupled-line lters and end-cou-
pled bandpass lters can be realized in CPW congura-
tions also [93,94]. Usually, the implementations of these
lters utilize a single-layer conguration that is much
easier to fabricate. Sometimes a double-layer geometry is
used for a wide bandwidth [95]. Inductively coupled band-
pass lters can also be realized using CPW. These lter
congurations eliminate the need for via holes to the
ground. The layout of an inductively end-coupled band-
pass lter in a CPW conguration is shown in Fig. 36.
8.4.3. Hybrid Ring Couplers and Ring Resonators. The
hybrid ring coupler can be realized in a CPWslotline con-
guration [96] as shown in Fig. 37a. This circuit makes
use of the properties of a slotline T junction and a CPW-
slotline T junction together with the usual operation of a
ratrace hybrid ring [97]. The circumference of the slot ring
is one wavelength divided equally into four quarter-wave
sections.
SS: Shorting strap
CPW
CPW
CPW
C
L
CPW
SS
CPW
SS
CPW
SS
L
C
Figure 35. Capacitor and inductor elements realized in CPW
conguration. C represents a capacitor and L represents an
inductor.
Figure 36. Layout of an inductively coupled multiresonator
bandpass lter realized in a CPW conguration.
(a)
CPW
Microstrip
Slotline
(b)
H E
1
2
/4
/4 /4
/4
Figure 37. (a) Layout of a hybrid ring coupler in a CPW cong-
uration where the slot ring is one wavelength (l) and divided
equally into four quarter-wave sections; (b) circuit conguration
of a slotline ring resonator with different coupling schemes such
as CPW, slotline, or microstrip line.
Ring resonators can also be realized using CPWor slot-
line feed as shown in Fig. 37b [98]. The resonator can be
easily integrated with shunt and series active devices and
used for several applications [99,100]. The CPW-fed slot-
line ring resonators show characteristics similar to those
of microstrip-fed slotline ring resonators.
8.4.4. Mixers. Most of the microwave mixers use
Schottky barrier diodes for mixing operation. A typical
mixer circuit consists of a mixer diode together with
coupling networks for feeding the RF signal and local os-
cillator (LO) power and for extracting the IF signal. One of
the commonly used mixers is the balanced mixer.
Balanced mixers realized in a CPW conguration [101]
use a balanced local oscillator input and an unbalanced
signal input. One of the mixer congurations is shown in
Fig. 38. Local-oscillator voltage is applied via the slotline,
and the signal is fed through the coaxial line. Connection
to the slotline is made by a small copper coaxial cable (not
shown) at right angles to the slot, while a coaxial line
connection to the CPW is made directly along as shown.
The IF connecting wires are brought through holes in the
substrate to mixer diodes.
8.4.5. Oscillators. A three-port MESFET oscillator de-
signed in a CPW conguration [102] is shown in Fig. 39.
The gate of the device is self-biased in order to minimize
the number of bias points. The source and the drain are
connected to 50 O CPW transmission lines.
8.4.6. Distributed Ampliers. A low-noise distributed
CPW amplier [103] is shown in Fig. 40. The distributed
amplier topology offers not only typically greater than
octave bandwidth, but also an excellent phase linearity
due to the transmission line characteristics inherent in its
topology. A distributed amplier is made up of a set of
cascaded devices, which act as shunt capacitances, con-
nected together by high-impedance transmission lines
that simulate inductances. The CPW congurations offer
simplied fabrication processing and hence lower cost
compared to similar microstrip ampliers.
Various CPW circuit examples mentioned above pro-
vide a sampling of microwave circuits that can be designed
in CPW congurations. Several other CPW circuit exam-
ples have been reported in the literature [104107].
9. LUMPED-ELEMENT CIRCUITS
Circuits employing lumped elements, such as inductors
and capacitors, are used extensively at lower frequencies.
Lumped elements, by definition, are much smaller than
the wavelength (ol/10) at the operating frequency and
exhibit small phase shift across any physical dimension.
Thus, the frequency limit for lumped elements is depen-
dent on the size of the element. With improving technol-
ogy and with the development of monolithic microwave
circuits, miniaturization of electronic elements becomes a
possibility, and lumped-element circuits are used up to
about 20 GHz. In millimeter wave monolithic circuits,
lumped resistors and MIM capacitors are commonly
used. Spiral inductors can be designed with self-resonance
frequencies up to 40 GHz. There are several advantages of
lumped-element circuits compared to distributed element
transmission-line circuits [108, p. 118]. First, lumped el-
ements have smaller frequency-dependence and are there-
fore good for wideband circuits. Second, the use of lumped
elements affords a considerable size reduction (a factor of
10 in area) compared with distributed element circuits in
microwave integrated circuits. Third, since the substrate
area required is smaller and many lumped-element MICs
can be processed simultaneously, the lumped elements are
less costly.
Signal
LO
Metalization
Coax
Capacitor CPW
Slotline
Schottky
diode
Through-
hole
Figure 38. A circuit arrangement for a double balanced mixer
using CPW where RF signal is fed through the coaxial line, LO
voltage is applied via the slotline, and IF signal is brought out
from the through-holes. (Based on Ref. 101.)
Mesfet
Drain
27 mm 15 mm
1
5

m
m
7 mm
2
4

m
m
7

m
m
Gate
S
o
u
r
c
e
Figure 39. The circuit layout of an oscillator using CPW cong-
urations where the gate of the MESFET is self-biased and is con-
nected to a resonant section of open-circuited transmission line.
The dashed line shows the possible position of optical waveguides
in a MachZehnde modulator. All CPW lines were designed to
have an impedance of 50O. (From Ref. 102, r IEEE, 1993,
reprinted with permission.)
9.1. Lumped-Element Components
9.1.1. Inductors. Depending on the value of the induc-
tance required, lumped inductors can be realized either as
straight narrow strips (ribbon inductors), as single-loop
inductors, or as multiturn spiral inductors:
Ribbon Inductors. A microstrip ribbon inductor and its
equivalent circuit are shown in Fig. 41. For short
lengths (ol
g
/4) the inductance L and shunt capaci-
tances C at the two ends can be calculated [35]:
L
Z
0
2pf
sin
2pl
l
g
_ _
18
C
1
2pfZ
0
tan
pl
l
g
_ _
19
A narrow strip with high Z
0
is needed to achieve a
high inductance value with low parasitic capacitance.
However, in practice the choice of the strip width is
determined by fabrication limits, by the direct-cur-
rent-carrying capacity, and by the high resistance of
very narrow strips. The strip length is limited simply
by the need to ensure a realistic and economical chip
size. The ribbon inductor is thus limited to values of
less than 1 nH, but is a relatively pure inductor
with low parasitic capacitances.
Loop Inductors. An example of a single-loop inductor is
shown in Fig. 42. Because of their inefcient use of
the chip area, loop inductors have been used very lit-
tle in MMICs. However, design information can be
found in a number of references [109111].
Spiral Inductors. Spiral inductors with multiple turns
are essential for inductance values above approxi-
mately 1 nH. There are two kinds of spiral inductors:
circular spiral and square spiral (see Fig. 43). Circu-
lar (approximately) spirals have a slightly better
quality factor at the cost of layout complexity and
have a less convenient shape for integration with
other components. So the square spirals are used
more often in MMICs. Design equations can be found
in Refs. 112115. The drawback of the spiral induc-
tors is that the need to connect the center turn back to
the outside circuit dictates that either airbridge or
dielectric crossovers must be used. There are a num-
ber of different solutions for this connection problem,
and these are illustrated in Fig. 44.
Stacked Spirals. Stacked spiral inductors [83] com-
prise a pair of interwound spirals placed on separate
metal layers. The major advantage is that the turns
are much more tightly packed than what normal pho-
tolithography and metal patterning would allow for a
single-layer spiral. Since the turns are separated ver-
tically, there is less capacitance between adjacent
turns than there would normally be with such small
gaps. However, since the lower metal thickness is
limited to 1 mm or less, stacked spirals have higher
series resistances. They are generally used at fre-
quencies below 5 GHz or so.
RF in
z
in
RF out
HEMT HEMT HEMT HEMT
90 jm 120 jm 120 jm 120 jm 90 jm
HEMT
z
out
v
g
v
d
Figure 40. The circuit schematic of a
220-GHz CPW distributed amplier using
high-electron-mobility transistors (HEMTs)
and CPW lines. (From Ref. 103, r IEEE,
1993, reprinted with permission.)
Parasitic
capacitance
to ground
L
Z
0
, l
(b)
(a)
C C
Parasitic
capacitance
to ground
Figure 41. The ribbon inductor consisting of a high-impedance
transmission line section: (a) physical layout; (b) equivalent
circuit.
Figure 42. Layout of a single-loop inductor (not used much be-
cause of inefcient use of substrate area).
9.1.2. Capacitors. Both overlay metalinsulatormetal
(MIM) capacitors and interdigital capacitors are used in
microwave circuits. Interdigital capacitors can be used for
values up to approximately 1 pF, above which their size
and the resulting distributed effects prevent their use.
Overlay capacitors are therefore needed for most of appli-
cations, such as direct-current blocking and decoupling,
where large capacitor values are required:
Overlay Capacitors. These consist of an MIM, with the
most common insulators being silicon nitride, silicon
dioxide, and polyimide. Silicon nitride is popular
since it has a fairly higher e
r
compared to silicon
dioxide and polyimide. The type of connection used
from the capacitors to the rest of the circuit depends
on whether an airbridge or a polyimide-based two-
metal level process is used. Three types of overlay
capacitors with connections are shown in Fig. 45. De-
sign equations for C and capacitor quality factor Q
can be found in Ref. 108, pp. 161165.
Interdigital Capacitors. A capacitor, simpler than the
MIM capacitors, is the single-layer interdigital ca-
pacitor. The interdigital capacitor consists of a num-
ber of interleaved microstrip ngers coupled together
and is fabricated in a single-layer structure. Its struc-
ture and equivalent circuit are shown in Fig. 46. The
maximum value of capacitance of an interdigital ca-
pacitor is limited by its physical size. It can be fabri-
cated with values of 0.1pF to 15 pF in a reasonable
size. Because of the long length of gap between lines,
large capacitors resonate at low frequency. Therefore
only small value capacitors (less than 2 pF) are prac-
tical at higher frequencies. Since interdigital capaci-
tors do not use a dielectric lm, their capacitance
tolerance is very good and is limited only by the ac-
curacy of the metal pattern definition. Hence, they
are ideal as tuning, coupling, and matching elements
where small capacitor values are required but precise
values are necessary. Design equations for these
capacitors are available in Refs. 116 and 117.
9.1.3. Resistors. Because of the signal loss associated
with resistances, the resistors are not used as much at
microwave frequencies as in the low-frequency circuits.
They are, however, irreplaceable in many cases such as for
reectionless terminations, for suppressing undesirable
signals, for damping elements providing the isolation of
circuits from each other, or for reducing the power level.
Resistors can either use doped semiconductor layers
(mesa resistors or implanted planar resistors) or sput-
tered thin-lm resistive layers. Figure 47 shows the two
Substrate
(b)
(a)
d
o
d
i
s
t
w
Figure 43. Spiral inductors: (a) circular spiral; (b) square spiral.
Square spirals allow a more compact layout of components in
MMICs.
Port 2 Port 1
Top plate
Bottom plate
(a)
(b)
(c)
Dielectric via
Port 2 Port 1
Port 2 Port 1
Figure 45. Overlay capacitors: (a) using a dielectric via; (b) with
an airbridge; (c) without an airbridge or spacer dielectric.
(a) (c)
(b) (d)
Figure 44. Different interconnection schemes for spiral induc-
tors: (a) single airbridge; (b) airbridges over an underpass;
(c) formed entirely of airbridges; (d) using two metal levels
for an underpass. (From Ref. 83, r IEE, 1995, reprinted with
permission.)
resistor examples. In either case, since the layer or lm
thickness is xed, it is very convenient to quote resistivity
in terms of an ohm-per-square gure. Hence, the value of
the resistor is chosen by selecting a suitable aspect ratio. A
practical limit is imposed by the higher parasitic capaci-
tance of large pads and the resistors physical size.
GaAs Mesa Resistors. The term mesa refers to the
conguration where the whole wafer is doped and
subsequently etched selectively so that active regions
remain only where required. Figure 48 shows the de-
tails of a mesa resistor with passive and negative
mesa edges [83, pp. 7778]. The GaAs resistor relies
on the linearity of the semiconductors current-eld
characteristic at low electric eld values. Hence it is
important to consider how much current is to be
passed through the resistor. The maximum electric
eld allowed in the resistor is normally given as a
volts per unit length value for a certain permissible
percentage deviation from perfect linearity.
Thin Film Resistors. Sputtered thin-lm resistors offer
improved linearity and lower temperature coef-
cients compared with the mesa types. In addition,
the ohm-per-square gure can be optimized for the
circuit designer without any limitations imposed by
the requirements of the active devices. The most com-
monly used materials are tantalum nitride, cermet,
and nickel chrome. Their temperature coefcients of
resistance are less than one-tenth that of GaAs, and
ohm-per-square gures of 50 O can be produced,
which is convenient for the circuit designers [118,
pp. 158159].
9.2. Examples of Lumped-Element Circuits
Planar Spiral Transformers. A lumped element trans-
former with spiral inductors offers direct-current
(DC) blocking, matching, and DC bias injection func-
tions in a very small size [120,121]. A planar spiral
transformer, whose layout and equivalent circuit are
shown in Fig. 49, is used as a lumped transformer up
to 4GHz or so [83, p. 73]. The two-coupled inductors
have self-inductance and mutual inductance. There
are series resistances in the conductors, interturn ca-
pacitance, and shunt capacitance to ground. The most
serious parasitic is usually the capacitance between
the two spirals since this makes the transformer res-
onant as the capacitive coupling becomes dominant at
higher frequencies. To minimize this parasitic capac-
itance while achieving a high mutual inductance, the
turns need to be very narrow and close together. Typ-
ically a transformer would have 5mm conductor
widths and 5mm gaps.
Lumped-Element Matching Circuits for MMIC Ampli-
ers. A lumped-element amplier circuit [83, p. 19]
is shown in Fig. 50. Lumped-element matching net-
works (using spiral inductors are overlay capacitors)
provide an appropriate arrangement at frequencies
below 20 GHz. The chip is a 12 GHz single-stage am-
plier. When operation frequency is higher than
20 GHz, the performance of spiral inductors degrades
because of their self-reasonances.
Terminal
strip
Low-frequency equivalent circuit
2 1
R series C
2
C
T
C
T
C
1
C
1
N fingers
Unit
cell
X
2
X
Substrate
Ground
plane
Q =
1
C
2
R
4 X
R = R
S
4
3
l
XN
r
W
l
c
Figure 46. Interdigital capacitor layout and equivalent circuit.
N is the number of ngers, 4X is the width of a unit cell. (From
Ref. 116, r IEEE, 1970, reprinted with permission.)
Contact
pad
(a) (b)
Trim mask
(if applicable)
Mass
Figure 47. Lumped resistor examples: (a) small value ( 50O);
(b) large value ( 3000O). (From Ref. 83, rIEE, 1995, reprinted
with permission.)
Etched
surface
of S.I.
substrate
Ohmic
contact
metal
Negative
mesa slope
Positive
mesa slope
Figure 48. Mesa resistor view showing positive and negative
mesa edges.
Lumped-Element Resonators. Series and parallel
lumped-element resonators using interdigital capac-
itors are shown in Fig. 51 [117]. These kinds of res-
onators can be used at frequencies of 7 GHz and
higher. A 0.3-pF capacitor with a 20 mm gap had a
capacitance tolerance of less than 25% from sample to
sample. These resonators are fabricated on one side of
alumina or quartz substrate.
Filters. Figure 52 depicts a lumped element lter with
bandstop loss of 30 dB at 9 GHz (117). The equivalent
circuit is also shown. Interdigital capacitors are of
practical use and allow the attainment of the few
picofarad capacitor values required for the design of
these lters at the higher microwave frequencies.
Several other types of lumped element microwave
circuits have been reported in the literature
[83,118,119,121].
10. MULTILAYER MICROWAVE CIRCUITS
10.1. Multilayer Congurations
Portable microwave systems require circuits to be fabri-
cated in smaller sizes and volumes. Multilayer congura-
tions have begun to be adapted for microwave circuits to
meet these size requirements [122125]. Also, multilayer
designs exhibit more exibility and, in several cases, yield
better performance than the corresponding designs in sin-
gle-layer congurations [126]. A tight coupling is an out-
standing characteristic that multilayer circuits can
provide. Since a physical geometry with very narrow spac-
ing is necessary to produce a tight coupling in single-layer
directional couplers, directional couplers realized in two
layers are well suitable for providing a tight coupling and
high directivity [127]. This tight coupling also makes it
possible to obtain wide bandwidth in lter circuits [123].
In addition, the circuits designed using symmetrical pla-
nar structures in traditional single-layer congurations
can now be implemented in asymmetrical geometry. This
asymmetry produces more design exibility, leading to
conveniently realizable physical geometries.
Multilayer congurations have also been utilized to in-
tegrate a number of passive components and active devic-
es into a module to reduce the size and the volume of the
whole system. This kind of use of multilayer structures
has resulted in the development of multichip module
(MCM) technology at microwave frequency [128,129].
However, multilayer congurations inherently require
C
3
L
2
C
2
C
1
R
1
R
2
L
1
M
(b)
(a)
Figure 49. A lumped-element spiral transformer: (a) layout;
(b) equivalent circuits. (From Ref. 83, r IEE, 1995, reprinted
with permission.)
Input
V
D
V
D
V
G
V
G
Output
(a)
(b)
I
A
N
K
C
L
Output Input
Figure 50. A lumped-element 12-GHz MMIC amplier: (a) lay-
out; (b) circuit diagram. (From Ref. 83, r IEE, 1995, reprinted
with permission.)
vertical interconnections between adjacent layers. Also,
there is a possibility of air gap and misalignment between
different dielectric layers, and these can affect the circuit
performance.
10.2. Multilayer Microwave Circuit Technologies
10.2.1. Multilayer MMICs. Multilayer MMICs [130
132] are constructed using multilayer passive circuits in-
tegrated with active devices on the semiconductor sub-
strate. Active devices, resistors, and MIM capacitors are
formed on the surface of a semiconductor wafer. Dielectric
lms and conductors are stacked on the wafer, and trans-
mission lines and the ground layers are connected through
via holes. This structure allows transmission lines with
reduced linewidths, vertical interconnections with short
signal delays, and miniaturized connections in a small
area. In addition, this provides miniature but low-loss
transmission lines and increased design exibility.
For examples using this multilayer structure, several
miniature passive circuits such as directional couplers,
Wilkinson power dividers, transmission lines, and planar
baluns have been designed and fabricated. Active circuits
such as mixers, ampliers, phase shifters, and upconvert-
ers are integrated with passive circuits in planar forms.
Figure 53 [131] shows the typical conguration of a mul-
tilayer MMIC. In Fig. 53, a three-dimensional MMIC fab-
ricated on a GaAs substrate integrates active devices,
resistors, and MIM capacitors on the surface of a semi-
conductor wafer. A thin-lm microstrip (TFMS) line offers
a compact meanderline conguration while thin polyimide
lms and conductors are stacked on the wafer and ground
metal is inserted between layers.
10.2.2. Multichip Modules. A multichip module (MCM)
[129] is dened as multilayer sandwiches of dielectric and
conducting layers, on which integrated circuits and pas-
sive components (if any) are mounted directly on (or inside
of) the sandwich structure, without separate packaging
for each of the active components. That is, the chips are
mounted bare onto the MCMs, which then provide the
Limit of ground plane
3.5 mm
L = 1.6 nH
C = 0.2 pF
R = 2
5.7 mm
Output
50 microstrip
Input
L R
R
C
L
C
Figure 52. Bandstop lter and equivalent circuit. Bandstop loss
at 9GHz was 30dB. (FromRef. 117, rIEEE, 1971, reprinted with
permission.)
Narrow line width
(1030 m)
Ground and
shielding
metal
Polyimide
Microwire
Broadside
coupler
Spiral-/meander-like
TFMS line
S, I, GaAs
10 m
h
GND
MIM FET R
Narrow spacing
(3h < 30 m)
Figure 53. Structure of the three-dimensional multilayer MMIC
fabricated on GaAs substrate. Active devices and an MIM capac-
itor are placed on the surface of a wafer, thin polyimide lms and
conductors are stacked on the wafer, and thin-lm microstrip
(TFMS) lines and ground layers are connected through via holes.
(From Ref. 131, r IEEE, 1996, reprinted with permission.)
A
(a)
(b)
B
B
A
A
B
B
A
Figure 51. Series and parallel lumped-element resonator net-
works using interdigital capacitors: (a) series LC; (b) parallel LC.
The size AA
0
BB
0
ts across a standard coaxial connector. (From
required power and ground, as well as all the signal in-
terconnect and the electrical interface to the external en-
vironment. The entire MCM, including chips and passive
components, may be placed in a hermetic package much
like a large single-chip carrier, or it may be directly cov-
ered with a sealant material (such as epoxy or a glass
passivation coating) to protect the components from
physical damage.
Three general categories of MCMs are MCM-C (ceram-
ic), MCM-L (laminate), and MCM-D (deposited). MCM-Cs
are manufactured by stacking unred layers of ceramic
dielectric, onto which liquid metal lines are silk-
screened using a metal ink process. The individual inked
layers are then aligned, pressed together, and cored into
a solid planar structure, onto which integrated circuits
can be installed. MCM-Ls are manufactured through the
lamination of sheet layers of organic dielectric, and they
are very similar to traditional printed circuit-board tech-
nology. MCM-Ds are manufactured through the deposition
of organic or inorganic dielectrics onto a silicon or alumina
support substrate. After each dielectric layer is deposited,
one of several technique is used to pattern metal lines as
well as metal vias. The chips are then installed on the
upper surface.
MCM-D technology can provide a versatile platform for
the integration of GaAs MMICs and silicon devices for
microwave circuits where performance, size, and weight
are critical factors. Multilayer circuits such as spiral in-
ductors, baluns, directional couplers, Lange couplers,
transmission line transformers, lters, ampliers, and
voltage-controlled oscillators have been realized using
this technology [129]. Figure 54 shows a typical congu-
ration that can be fabricated by MCMD technology
suitable for design of multilayer microwave circuits.
10.3. Design of Multilayer Circuits
Multilayer microwave circuits can be divided into two
groups, the rst of which consists of planar components
employing multilayer transmission structures, while in
the second group, one integrates various circuits into a
single multilayer module.
Multilayer couplers, lters, baluns, inductors, and so
on, belong to the rst group because these employ multi-
layer transmission structures to overcome difculties oc-
curring in single-layer designs. It is well known [62] that
coupled lines realized in single-layer structures produce a
weak coupling. Thus, when a tight coupling is required,
coupled lines can be implemented in multilayer congu-
rations because overlapping geometry between coupled
lines is possible. Coupled-line couplers requiring a tight
coupling and those requiring high directivity are designed
[127] using this advantage. For the design of parallel cou-
pled-line lters and end-coupled lters, a wide bandwidth
is frequently required. For this purpose, multilayer struc-
tures used in coupled-line lters [133] and end-coupled
lters [95] allow overlapping geometry in the design,
which leads to a wide bandwidth and exibility in selec-
tion of physical dimensions. Baluns using multilayer cou-
pled lines [134] can also be designed for wideband
operation and with compact dimensions because a tight
coupling can easily be obtained in multilayer congura-
tions. Apertures in the ground plane can be employed to
achieve interconnection between different layers in mul-
tilayer circuits [135] such as aperture-coupled couplers
and magic tees.
For the multilayer circuits in the second group, passive
components and/or active devices are integrated into a
single module for miniaturization and cost-effective pro-
duction [131]. When these are integrated, vertical vias be-
tween different layers and apertures in the ground plane
are placed appropriately considering miniaturization,
crosstalk, and productivity.
10.4. Examples of Multilayer Circuits
Some examples of multilayer microwave circuits are dis-
cussed in this section.
10.4.1. Couplers. Since single-layer structure is not
convenient for couplers requiring a tight coupling, multi-
layer conguration has been explored to overcome this
difculty. As a result, coupled-line directional couplers in
two layers [136] and reentrant type couplers [137] de-
signed for a tight coupling have been reported. Figure 55
shows a two-layer coupler [136]. Two conductors with un-
equal widths are placed on different layers to yield a tight
coupling. Different port impedances can be used at the
four external ports.
Figure 56 shows a reentrant-type coupler where two
identical conductors placed on the top layer with a wider
conductor on the bottom layer produce a tight coupling.
The spacing between the two conductors on the top layer is
wide, not leading to tight tolerance requirement of a nar-
row gap in fabrication.
10.4.2. Filters. Multilayer congurations have been
used for wideband parallel coupled-line lters [138] and
SiO
2
Via stack Spiral inductor
Passivation
Polymide
Silicon/sapphire
Back metallization
Resistor
NiCr
Polymide
M4
M3
M2
M1
Figure 54. Multichip module (MCM-D) technology used for RF
integration. Polyimide dielectrics are deposited onto a silicon sap-
phire substrate; then the chips are installed on the upper surface,
and transmission lines and vias are fabricated.
4
Conductor 2
Conductor 1
2 1
3
Figure 55. Two-layer coupled-line coupler. Conductors are
placed on different levels. A tight coupling can be obtained.
end-coupled bandpass lters [95] because lters in single-
layer congurations cannot yield a wide bandwidth. Aside
from the wide bandwidth, enhanced freedom in circuit
layout in multilayer designs makes multilayer congura-
tions more attractive. Parallel coupled-line lters and end-
coupled bandpass lters have been developed in multilay-
er structure and also in multilayer coplanar waveguide
structure. Examples of these lter circuits are shown in
Figs. 57 and 58.
In Fig. 57, coupled lines, which can be placed on differ-
ent layers, constitute a bandpass lter where each coupled
line can have nonsymmetric geometry leading to exibil-
ity in design. Due to these coupled lines offering tight cou-
pling, multilayer coupled-line bandpass lters provide
wide bandwidths which are not achievable in single-lay-
er congurations. In Fig. 58, gap-coupled sections consti-
tute a bandpass lter in two-layer conguration. Since
this structure allows overlapping conductors in each gap-
coupled section, this can also be used for obtaining a wide
bandwidth.
10.4.3. Baluns. Due to the requirement for a wide
bandwidth and compact design, planar Marchand baluns
have been developed [134] using coupled lines in two-layer
congurations. These baluns provide wide bandwidth
(more than one octave) and compaction of physical dimen-
sions. Figure 59 shows one such circuit reported in the
literature. In Fig. 59, two kinds of coupled-line sections in
two layers constitute a planar Machand balun. An unbal-
anced signal entering port 1 is transformed into a bal-
anced signal coming out of ports 2 and 3.
Three-line baluns using three-coupled lines have also
been designed [139] in a two-layer structure, leading to
compact design and good performance. Figure 60 shows
one of such balun designs. In this design, one of the con-
ductors is placed on one layer while two other conductors
are placed on the second layer. With appropriate termi-
nations at the ports, this balun transforms an unbalanced
signal at port 1 to a balanced signal at ports 2 and 3.
10.4.4. Hybrids. Branchline couplers [140] and magic
tees [135] have been developed using multilayer congu-
rations to produce better performance than what can be
obtained in single-layer congurations. Figure 61 [140]
shows a multilayer branchline coupler where the two out-
put ports are mutually isolated, and the signals at these
two ports are out of phase by 901. The distance between
two microstrips or slotlines is one quarter-wavelength. It
is possible to make 3 dB branchline couplers by suitably
choosing the values of Z
m
and Z
s
.
Figure 62 shows [135] a multilayer magic tee where the
microstrip line (difference arm) on the lower substrate is
coupled to a slot-aperture and is terminated on the same
layer by an open-circuit stub. The other three magic tee
ports are located at the top layer. With the excitation at
port 1 (difference port), signals at ports 2 and 3 are equal
in magnitude and out of phase by 1801. When the sum port
(port 4) is fed, signals at ports 2 and 3 are in phase and
equal in magnitude. In addition to the circuits described
above, there are many other multilayer circuits like trans-
mission-line transformers [141], aperture-coupled multi-
layer circuits [142], and aperture-fed and microstrip patch
antennas [143].
11. DESIGN OF MICROWAVE CIRCUITS
As is true also for design processes in other domains of
engineering, the computer-aided design (CAD) approach
[144146] is being used extensively for the design of mi-
crowave circuits and is also being rened further. In order
Figure 57. Multilayer parallel coupled-line lter. Various con-
ductors can be placed on different levels; each coupled section
produces tight coupling, leading to broad bandwidth for the lter.
Cross-section at xx
x
x
Figure 58. Two-layer end-coupled bandpass lter. Half-wave
resonators shown in dotted lines are at lower layer. Each gap-
coupled section can have overlapping geometry for tight coupling
and therefore broad bandwidth of the lter.
Ground plane
Coupled lines B
Coupled lines A
1
2
3
Lower substrate
(GaAs)
Upper layer
(Si
3
N
4
)
Figure 59. Planar Marchand balun. Two kinds of coupled lines
constitute a balun transforming an unbalanced signal entering
into port 1 to a balanced signal coming out of ports 2 and 3. (From
W
1
W
1
S
3
4
2
1
W
2
Figure 56. Reentrant-type coupler. Two identical conductors are
placed on top layer with a wide spacing, a wide conductor is placed
on the bottom layer. This coupler provides a tight coupling.
to appreciate the CAD methodology, it is necessary to re-
view the conventional design process that designers used
before the CAD methods and software were developed.
11.1. Conventional Design Procedure
A ow diagram depicting the conventional design proce-
dure is shown in Fig. 63. One starts with the desired cir-
cuit specications and arrives at an initial circuit
conguration. Available design data and previous experi-
ence are helpful in selecting this initial conguration.
Analysis and synthesis procedures are used for deciding
values of various parameters of the circuit. A laboratory
model is constructed for the initial design, and measure-
ments are carried out for evaluating its characteristics.
Performance achieved is compared with the desired
specications; if the given specications are not met, the
circuit is modied. Adjustment, tuning, and trimming
mechanisms incorporated in the circuit are used for car-
rying out these modications. Measurements are carried
out again and the results are compared with the desired
specications. The sequence of modications, measure-
ments, and comparison is carried out iteratively until
the desired specications are achieved. At times the spec-
ications are compromised in view of the practically
feasible performance of the circuit. The nal circuit
conguration thus obtained is sent for prototype fabrica-
tion.
The above procedure had been used for the design
of microwave circuits for quite some time. However,
it became increasingly difcult to use this iterative
experimental method successfully because of the follow-
ing considerations:
1. Increased complexity of modern systems demands
more precise and accurate design of circuits and
subsystems. Consequently, the effect of tolerances in
the circuit design becomes increasingly important.
2. A larger variety of active and passive components,
as described in previous sections, are now available
for achieving a given circuit function. The choice of
the appropriate device or transmission structure
becomes difcult if the iterative experimental ap-
proach is used.
Port 3
Slot
Ground
Port 2
Port 1
Port 4
Figure 62. A two-layer microstrip magic-tee junction (ports 2, 3,
and 4) coupled through a slot in the ground to a microstrip line
(port 1) on the backside. (From Ref. 135, rIEEE, 1997, reprinted
with permission.)
Circuit
specifications
Design
data
Final
fabrication
Modifications
Fail
Pass
Compare
Initial
circuit
design
Laboratory
model
Measurements
Figure 63. The conventional design procedure that was used for
microwave circuits before CAD methods were developed.
Microstrip
line
Slotline
#3
#2
#4
#1
2L
s
2 x
y
2L
m
z
m
z
s
w
m
w
s
2L
s
1
Figure 61. A branchline coupler in a two-layer structures. Slot-
lines are placed on the lower layer while two microstrip lines are
located on the upper layer, and two outputs are (at ports 2 and 3)
901 out of phase and equal in magnitude. (FromRef. 140, rIEEE,
1995, reprinted with permission.)
w
1
w
1
w
2
h
2
h
1
s
In(1)
Out(3) Out(2)
r1
r2
Figure 60. Two-layer three-line balun. One conductor is placed
on one layer while two other conductors are placed on the other
layer. With appropriate terminations at different ports, this cir-
cuit works as a balun.
3. It is very difcult to incorporate any modications in
the circuits fabricated by MIC technology.
The method developed for dealing with this situation is
known as computer-aided design (CAD). Computer-aid-
ed design in its strict interpretation may be taken to mean
any design process where the computer is used as a tool.
However, usually the term CAD implies that without the
computer as a tool, that particular design process would
have been impossible or much more difcult, more expen-
sive, more time-consuming, and less reliable and more
than likely would have resulted in an inferior product.
11.2. CAD of Microwave Circuits
A typical ow diagram for the CAD procedure is shown in
Fig. 64. As before, one starts with a given set of specica-
tions. Synthesis methods and available design data (at
times prestored in computer memory) help to arrive at the
initial circuit design. The performance of this initial cir-
cuit is evaluated by a computer-aided circuit analysis.
Numerical models for various components (passive and
active) used in the circuit are needed for the analysis.
These are called from the library of subroutines developed
for this purpose. Circuit characteristics obtained as re-
sults of the analysis are compared with the given speci-
cations. If the results fail to satisfy the desired
specications, the designable parameters of the circuit
are altered in a systematic manner. This constitutes the
key step in the optimization. Several optimization strate-
gies include sensitivity analysis of the circuit for calculat-
ing changes in the circuit parameters. The sequence of
circuit analysis, comparison with the desired perfor-
mance, and parameter modication is performed itera-
tively until the specications are met or the optimum
performance of the circuit (within the given constraints) is
achieved. The circuit is now fabricated and the experi-
mental measurements are carried out. Some modications
may still be required if the modeling and/or analysis are
not accurate enough. However, these modications, hope-
fully, are very small, and the aim of the CAD method is to
minimize the experimental iterations as far as practicable.
The process of CAD, as outlined above, consists of three
important segments:
1. Modeling
2. Analysis
3. Optimization
Modeling involves characterization of various active and
passive components to the extent of providing a numerical
model that can be handled by the computer. In the case of
microwave circuits, one comes across a variety of active
and passive elements. Semiconductor devices used include
bipolar and MESFET transistors, point contact and
Schottky barrier detectors, varactor and pin diodes, and
transferred electron and avalanche devices. Passive ele-
ments used in microwave circuits include sections of var-
ious transmission structures, lumped components, YIG
and dielectric resonators, nonreciprocal components, and
planar (two-dimensional) circuit elements. Transmission
structures could be coaxial line, waveguide, stripline, mi-
crostrip line, coplanar line, slotline, or a combination of
these. As mentioned earlier in article, not only do these
transmission structures need to be characterized fully
for impedance, phase velocity, and so on, it also becomes
necessary to model the parasitic reactances caused by geo-
metrical discontinuities in these densely packed transmis-
sion lines.
Modeling of components in microwave circuits had
been the main difculty in successful implementation of
CAD techniques at microwave frequency. However, the
development of electromagnetic (EM) simulation tech-
niques developed since the late 1980s has helped to con-
struct adequate models and bring microwave hybrid and
monolithic circuit CAD software to level of maturity. Mod-
eling still remains the major bottleneck for CAD of certain
classes of microwave circuits [such as coplanar waveguide
(CPW) circuits, multilayered circuits, and integrated cir-
cuit-antenna modules] and most of the millimeter-wave
(above B40 GHz) circuits. Current research in efcient
use of EM simulation techniques and in use of articial
neural network models [77] will lead to further improve-
ment in CAD tools for microwave and millimeter-wave
circuits.
Circuit
specifications
Sensitivity
analysis
Compare
Compare
Fabrication
Modify
Measurements
Fail
Pass
Pass
Fail
Models
Design data
and synthesis
methods
Initial
circuit
design
Analysis Modifications
Figure 64. Computer-aided design methodology used for micro-
wave circuits. Recent developments have brought microwave
CAD tools to a level of maturity.
12. EMERGING TRENDS IN MICROWAVE CIRCUITS
Even after going through developments over almost a cen-
tury, the eld of microwave circuits continues to see de-
velopments in full swing even today; particularly in the
areas of multilayered circuits, integrated circuitantenna
modules, quasioptical circuits and systems, and exotic
methodologies for CAD.
12.1. Multilayered Circuits
Microwave multilayered circuits were described earlier in
Section 10. Multilayered circuit technology is responsible
for two novel innovations in monolithic microwave cir-
cuits. The rst one of these is the development of three-
dimensional passive circuit technology [147] as an evolu-
tion from the two-dimensional planar microwave circuit
technology, which has been the cornerstone of integrated
microwave circuit technology so far. Three-dimensional
microwave circuits make use of multilayered congura-
tions combined with vertical wall-like microwave struc-
tures. This technology can reduce the sizes of inductors
and transmission lines to one-third or one-fourth of the
size of two-dimensional congurations.
Another novel concept that has emerged from the ap-
plication of multilayer circuit technique is that of master
slice MMIC [148]. The basic idea is similar to that of gate
arrays used for semiconductor application-specific inte-
grated circuits (ASICs). In an MMIC master slice, many
units that contain transistors, resistors, and lower elec-
trodes for MIM capacitors are located repeatedly (nearly
6 units/mm
2
) on a GaAs or Si wafer to form a master array.
Additional upper layers provide custom design intercon-
nects and top conductor for capacitors to yield semi-cus-
tom-designed circuits.
12.2. Integrated Circuit-Antenna Modules
Another very interesting trend in microwave and milli-
meter-wave design is the integration of conventional cir-
cuit and antenna functions in single components. This
integration has been made possible by common fabrication
technology used for planar circuits and printed antennas.
Various design approaches applicable to integrated cir-
cuitantenna modules have been developed [149]. Quasi-
optical systems [150] designed to generate high power at
millimeter-wave frequencies by employing a grid array of
sources are important examples of integrated circuitan-
tenna modules.
12.3. Knowledge-Aided Design (KAD)
Emerging innovations in design techniques [77] for mi-
crowave circuits include use of articial neural networks
and development of knowledge-based design tools.
Articial neural networks (ANNs) are neuroscience-in-
spired computational tools that learn from experience
(training), generalize from previous examples to new
ones, and abstract essential characteristics from inputs
containing noise or irrelevant data. A significant applica-
tion of ANN computing to microwave design has been for
development of component models [151]. These models are
as accurate as results for EM simulators used for training
them but are as efcient as network models. ANNs are
expected to play a significant role in microwave design in
areas of modeling and optimization and possibly as a
means of embedding knowledge in design tools.
Any design process involves several steps starting from
problem identication and going through specication
generation, concept generation, initial analysis and eval-
uation, and initial design to a detailed optimized design. It
is for the last of these steps (from initial design to detailed
optimized design) that the current microwave CAD tools
have been developed. The earlier steps in the design pro-
cess are not at all trivial. We need a technology to aid the
earlier stages of design where designers make important
and expensive decisions. Knowledge-based system seems
to be the most appropriate technology for this purpose.
BIBLIOGRAPHY
1. G. C. Southworth, Hyper-frequency waveguidesgeneral
considerations and experimental results, Bell Syst. Tech. J.
15:284309 (1936).
2. J. R. Carson, S. P. Meade, and S. A. Schelkunoff, Hyper-fre-
quency waveguidesmathematical theory, Bell Syst. Tech.
J. 15:310333 (1936).
3. G. C. Southworth, Some fundamental experiments with
waveguides, Proc. IRE 25:807822 (1937).
4. G. C. Southworth, Survey and history of the progress of
microwave arts, Proc. IRE 50:11991206 (1962).
5. P. H. Smith, Transmission line calculator, Electronics 12:29
31 (1939) An improved transmission line calculator, ibid.,
17:130133, 318, 320, 322, 324325 (1994).
6. A. G. Fox, An adjustable waveguide phase changer, Proc.
IRE 35:14891498 (1947).
7. W. A. Tyrell, Hybrid circuits for microwaves, Proc. IRE
35:12941306 (1947).
8. W. W. Mumford, Directional couplers, Proc. IRE 35:160165
(1947).
9. R. M. Barrett and M. H. Barnes, Microwave printed circuits,
Natl. Conf. Airborne Electronics, IRE, Dayton, OH, 1951.
10. R. W. Peters et al., Handbook of Tri-plate Microwave Com-
ponents, Sanders, Nashua, NH, 1956.
11. H. Howe, Jr., Stripline Circuit Design, Artech House, Ded-
ham, MA, 1974.
12. H. A. Wheeler, Directional Coupler, U.S. Patent 1,606,974
(1952).
13. D. D. Greig and H. F. Engelmann, Microstripa new trans-
mission technique for the kilomegacycle range, Proc. IRE
40:16441650 (1952).
14. F. Assadourian and E. Rimai, Simplied theory of microstrip
transmission systems, Proc. IRE 40:16511657 (1952).
15. H. A. Wheeler, Transmission line properties of parallel wide
strips by conformal mapping approximation, IEEE Trans.
Microwave Theory Tech. MTT-12:280289 (1964).
16. H. A. Wheeler, Transmission line properties of parallel strips
separated by a dielectric sheet, IEEE Trans. Microwave The-
ory Tech. MTT-13:172185 (1965).
17. S. W. Schilling and R. K. Durnwirth, The real world of mi-
cromin substrates, Parts 15, Microwave, 7(12):5256
(1968); 8(1):4446 (1969); 8(3):5759 (1969); 8(9):3641
(1969); 8(12):5457 (1969).
18. Microwave integrated circuits (special issue), IEEE Trans.
Electron. Devices ED-15 (1968).
19. Microwave integrated circuits (special issue), IEEE Trans.
Microwave Theory Tech. MTT-19 (1971).
20. K. C. Gupta and A. Singh, eds., Microwave Integrated Cir-
cuits, Halsted Press, New York, 1974.
21. J. Frey, ed., Microwave Integrated Circuits, Artech House,
Dedham, MA, 1974.
22. S. B. Cohn, Slot line on a dielectric substrate, IEEE Trans.
23. K. C. Gupta et al., Microstrip Lines and Slotlines, 2nd ed.,
Artech House, Norwood, MA, 1996.
24. C. S. Aitchison et al., Lumped microwave circuits, Parts IV,
Des. Electron. 8(11):2328 (1971); 9(1):3039 (1971); 9(2):42
51 (1971); 9(3):47 (1971); 9(4):4148 (1972); also Philips
Tech. Rev. 32:305314 (1971).
25. R. S. Pengelly and D. C. Rickard, Design, measurements and
application of lumped elements up to J-band, Proc. 7th Eur.
Microwave Conf., Copenhagen, 1977, pp. 460464.
26. T. Okoshi and T. Miyoshi, The planar circuitan approach
to microwave integrated circuitry, IEEE Trans. Microwave
Theory Tech., MTT-20:245252, 1972.
27. E. W. Mehal and R. W. Wacker, GaAs integrated microwave
circuits, IEEE Trans. Electron. Devices ED-15:513516
(1968).
28. M. M. Hasan and S. K. Mullick, Monolithic MICs, in K. C.
Gupta and A. Singh, eds., Microwave Integrated Circuits,
Halsted Press, New York, 1974, pp. 259277.
29. C. A. Liechti, Future of microwaves is monolithic, Microwave
Syst. News 8(11):60 (1978).
30. Microwave eld-effect transistors (special issue), IEEE
Trans. Microwave Theory Tech. MTT-24 (1976).
31. K. C. Gupta, Microwaves, Halsted Press, New York, 1980,
Chap. 10.
32. B. S. Perlman and V. G. Glenovatch, in L. Young and H.
Sobol, eds., Computer Aided Design, Simulation and Opti-
mization in Advances in Microwaves, Vol. 8, Academic Press,
New York, 1974, pp. 321399.
33. J. F. White, Semiconductor Control, Artech House, Dedham,
MA, 1977, pp. 177243.
34. D. M. Pozar, Microwave Engineering, Addison-Wesley, Read-
ing, MA, 1990.
35. T. Edwards, Foundations for Microstrip Circuit Design, 2nd
ed., Wiley, New York, 1992.
36. D. Fisher and I. Bahl, Gallium Arsendie IC Applications
Handbook, Vol. 1, Academic Press, San Diego, 1995.
37. T. K. Ishii, Handbook of Microwave Technology, Vol. 1: Com-
ponents and Devices, Academic Press, San Diego, 1995.
38. J. Uher, J. Bornemann, and U. Rosenberg, Waveguide Com-
ponents for Antenna Feed Systems: Theory and CAD, Artech
House, Dedham, MA, 1993.
39. J. J. Izadian and S. M. Izadian, Microwave Transition De-
sign, Artech House, Dedham, MA, 1988, Chaps. 3 and 4.
40. H. M. Barlow, The relative power-carrying capacity of high-
frequency waveguides, IEE Proc. 99(Part III):21 (1952).
41. R. E. Collin, Waveguide phase changer, Wireless Eng. 32:82
88 (1955).
42. R. Chatterjee, Elements of Microwave Engineering, Wiley,
New York, 1986.
43. R. Vahldieck, Quasi-planar lters for millimeter-wave appli-
cations, IEEE Trans. Microwave Theory Tech. 37:324334
(1989).
44. P. J. Meier, Integrated nline millimeter-wave components,
IEEE Trans. Microwave Theory Tech. MTT-22:12091219
(1974).
45. L. Q. Bui, D. Ball, and T. Itoh, Broad-band millimeter-wave
E-plane bandpass lters, IEEE Trans. Microwave Theory
Tech. MTT-32:16551658 (1984).
46. Y. Konishi and K. Uenakada, The design of bandpass lters
with inductive strip-planar circuit mounted in waveguide,
(1974).
47. W. Hauth, R. Keller, and U. Rosenberg, The corrugated-
waveguide band-pass ltera new type of waveguide lter,
Proc. 18th European Microwave Conf., Stockholm, Sweden,
1988, pp. 945949.
48. R. Levy, Synthesis of high power harmonic rejection wave-
guide lters, IEEE MTT-S Int. Microwave Symp. Digest,
1969, pp. 286290.
49. R. Levy, Tapered corrugated waveguide low-pass lters,
(1973).
50. G. F. Craven and C. K. Mok, The design of evanescent mode
waveguide bandpass lters for a prescribed insertion loss
characteristics, IEEE Trans. Microwave Theory Tech. MTT-
19:295308 (1971).
51. R. V. Snyder, Newapplication of evanescent mode waveguide
to lter design, IEEE Trans. Microwave Theory Tech. MTT-
25:10131021 (1977).
52. J. Bornemann and F. Arndt, Transverse resonance, standing
wave, and resonator formulations of the ridge waveguide
eigenvalue problem and its application to the design of E-
plane nned waveguide lters, IEEE Trans. Microwave The-
ory Tech. MTT-38:11041113 (1990).
53. R. E. Collin, Foundations for Microwave Engineering, Mc-
Graw-Hill, New York, 1992.
54. N. Marcuvitz, Waveguide Handbook, Dover, New York,
1965.
55. A. Weisshaar, M. Mongiardo, and V. Tripathi, Full-wave
computer-aided design of waveguide components by circuit
simulation, Int. J. RF Microwave Comput. Aided Eng.
8:236247 (1998).
56. W. K. Gwarek, Analysis of an arbitrarily shaped planar cir-
cuita time domain approach, IEEE Trans. Microwave The-
ory Tech. MTT-33:10671072 (1985).
57. R. L. Ferrari, Finite element analysis of three-dimensional
electromagnetic devices, Proc. 15th European Microwave
Conf., Paris, France, 1985, pp. 10641069.
58. W. J. R. Hoefer, The transmission-line matrix method: The-
ory and applications, IEEE Trans. Microwave Theory Tech.
MTT-23:882893 (1985).
59. P. Clarricoats et al., Numerical solution of waveguide dis-
continuity problems, IEE Proc. 114:878886 (1967).
60. T. Vasilyeva et al., Computer-aided analysis of slant dielec-
tric E-plane interface discontinuity in a rectangular wave-
guide, Int. J. RF Microwave Comput. Aided Eng. 8:248255
(1998).
61. F. Alessandri et al., Efcient full-wave automated design and
yield analysis of waveguide components, Int. J. RF Micro-
wave Comput. Aided Eng. 8:200207 (1998).
62. G. Matthaei, L. Young, and E. M. T. Jones, Microwave Fil-
ters, Impedance-Matching Networks, and Coupling Struc-
tures, Artech House, Dedham, MA, 1980.
63. G. L. Ragan, Microwave Transmission Circuits, Radiation
Laboratory Series 9, Boston Tech., Lexington, MA, 1964.
64. J. H. Bryant, Coaxial transmission lines, related two-con-
ductor transmission lines, connectors, and components: A
U.S historical perspective, IEEE Trans. Microwave Theory
Tech. MTT-32:970983 (1984).
65. B. Bhat and S. K. Koul, Stripline-Like Transmission Lines
for Microwave Integrated Circuits, Wiley, New York, 1989,
pp. 566687.
66. S. B. Cohn, Characteristic impedance of the shielded strip
transmission line, IEEE Trans. Microwave Theory Tech.
MTT-2:5257 (1954).
67. B. E. Spielman, Dissipation loss effects in isolated and cou-
pled transmission lines, IEEE Trans. Microwave Theory
Tech. MTT-25:648656 (1977).
68. R. Levy and L. F. Lind, Synthesis of symmetrical branch
guide directional couplers, IEEE Trans. Microwave Theory
Tech. MTT-16:8089 (1968).
69. J. K. Shimizu and E. M. T. Jones, Coupled transmission line
directional couplers, IEEE Trans. Microwave Theory Tech.
MTT-6:403410 (1958).
70. C. Y. Pon, Hybrid ring directional coupler for arbitrary pow-
er divisions, IEEE Trans. Microwave Theory Tech. MTT-
9:529535 (1961).
71. E. J. Wilkinson, An N-way hybrid power divider, IEEE
Trans. Microwave Theory Tech. MTT-8:116118 (1960).
72. S. B. Cohn, Parallel-coupled transmission line resonator l-
ters, IEEE Trans. Microwave Theory Tech. MTT-6:223231
(1958).
73. R. V. Garver, Diode switching, Microwave Syst. News
13:194204 (1983).
74. J. F. White, Diode phase shifters for array antennas, IEEE
Trans. Microwave Theory Tech. MTT-22:658674 (1974).
75. M. V. Schneider, Microstrip lines for microwave integrated
circuits, Bell Syst. Tech. J. 48:14211444 (1969).
76. R. K. Hoffmann, Handbook of Microwave Integrated Cir-
cuits, Artech House, Norwood, MA, 1987.
77. K. C. Gupta, Emerging trends in millimeter-wave CAD,
IEEE Trans. Microwave Theory Tech. 46:747755 (1998).
78. G. Gonzales, Microwave Transistor Ampliers Analysis and
Design, 2nd ed., Prentice-Hall, Upper Saddle River, NJ,
1997.
79. A. Sweet, MIC and MMIC Amplier and Oscillator Circuit
Design, Artech House, Dedham, MA, 1990.
80. F. Filicori, V. A. Monaco, and G. Vannini, A design method
for parallel feedback dielectric resonator oscillators, Proc.
19th European Microwave Conf., London, 1989, pp. 412417.
81. C. Y. Chang and T. Itoh, A varactor-tuned active microwave
band-pass lter, IEEE MTT-S Int. Microwave Symp., 1990,
pp. 499502.
82. J. A. Coekin, High-Speed Pulse Techniques, Pergamon, Ox-
ford, UK, 1975.
83. I. D. Robertson, MMIC Design, Inst. Electr. Eng., London,
1995.
84. C. P. Wen, Coplanar waveguide: A surface strip transmission
line suitable for non-reciprocal gyromagnetic device appli-
cation, IEEE Trans. Microwave Theory Tech. MTT-17:1087
1090 (1969).
85. J. B. Knorr and K. D. Kuchler, Analysis of coupled slots and
coplanar strips on dielectric substrate, IEEE Trans. Micro-
wave Theory Tech. MTT-23:541548 (1975).
86. R. W. Jackson, Considerations in the use of coplanar wave-
guide for millimeter-wave integrated circuits, IEEE Trans.
87. A. Wexler, Solution of waveguide discontinuities by modal
analysis, IEEE Trans. Microwave Theory Tech. MTT-
15:508517 (1967).
88. M. Naghed and I. Wolff, A three-dimensional nite-differ-
ence calculation of equivalent capacitors of coplanar wave-
guide discontinuities, IEEE MTT-S Int. Microwave Symp.,
1990, pp. 11431146.
89. T. W. Huang and T. Itoh, Full-wave analysis of cascaded
junction discontinuities of shielded coplanar type transmis-
sion lines considering the nite metallization thickness ef-
fect, IEEE MTT-S Int. Microwave Symp., 1992, pp. 995998.
90. H. Jin and R. Vahldieck, Full-wave analysis of coplanar
waveguide discontinuities using the frequency domain TLM
method, IEEE Trans. Microwave Theory Tech. 41:15381542
(1993).
91. R. W. Jackson, Mode conversion at discontinuities in nite-
width conductor-backed coplanar waveguide, IEEE Trans.
Microwave Theory Tech. 37:15821589 (1989).
92. &lts;o4S. B. Worm, Full-wave analysis of discontinuities in
planar waveguides by the method of lines using a source
approach, IEEE Trans. Microwave Theory Tech. MTT-
38:15101514 (1990).
93. F. Mernyei, I. Aoki, and H. Matsuura, MMIC bandpass lter
using parallel coupled CPW lines, Electron. Lett. 30:1862
1863 (1983).
94. D. F. Williams and S. E. Schwarz, Design and performance of
coplanar waveguide bandpass lters, IEEE Trans. Micro-
wave Theory Tech. MTT-31:558566 (1983).
95. W. Schwab, F. Boegelsack, and W. Menzel, Multilayer sus-
pended stripline and coplanar line lters, IEEE Trans. Mi-
crowave Theory Tech. MTT-42:14031407 (1994).
96. C.-H. Ho, L. Fan, and K. Chang, Broad-band uniplanar hy-
brid-ring and branch-line couplers, IEEE Trans. Microwave
Theory Tech. MTT-41:21162125 (1993).
97. P. Troughten, High Q-factor resonator in microstrip, Elec-
tron. Lett. 4:520522(1968).
98. C.-H. Ho, L. Fan, and K. Chang, Slotline annular ring ele-
ments and their applications to resonator, lter and coupler
design, IEEE Trans. Microwave Theory Tech. MTT-41:1648
1653 (1993).
99. K. Chang et al., On the study of microstrip ring and va-
racter-turned ring circuits, IEEE Trans. Microwave Theory
Tech. MTT-35:12881295 (1987).
100. J. A. Navarro, L. Fan, and K. Chang, The coplanar wave-
guide-fed electronically tunable slotline ring resonator,
IEEE MTT-S Int. Microwave Symp., 1992, pp. 951954.
101. J. K. Hunton, A microwave integrated circuit balanced mix-
er with broad-bandwidth, Proc. Microelectronics Symp.,
1969, pp. A3.1A3.2.
102. V. Radisic, V. Jevremovic, and Z. B. Popovic, CPW oscillator
conguration for an electrooptic modulator, IEEE Trans.
Microwave Theory Tech. 41:16451647 (1993).
103. K. Minot, B. Nelson, and W. Jones, A low noise, phase linear
distributed coplanar waveguide amplier, IEEE Trans. Mi-
crowave Theory Tech. 41:16501653 (1993).
104. L. Fan and K. Chang, Uniplanar power dividers using cou-
pled CPW and asymmetrical CPS for MICs and MMICs,
IEEE Trans. Microwave Theory Tech. 44:24112420 (1996).
105. L. Giauffret, J.-M. Laheurte, and A. Papiernik, Study of var-
ious shapes of the coupling slot in CPW-fed microstrip an-
tennas, IEEE Trans. Anten. Propag. Soc. 45:642647 (1997).
106. V. Trifunovic and B. Jokanovic, Review of printed Marchand
and double Y baluns: Characteristics and application,
(1994).
107. G. Forma and J. M. Laheurte, CPW-fed oscillating micro-
strip antennas, IEEE Antennas Propag. Soc. Int. Symp.,
1996, pp. 526529.
108. I. Kneppo and J. Fabian, Microwave Integrated Circuits,
Chapman & Hall, London, 1994.
109. R. S. Pengelly, Microwave Field Effect Transistors: Theory,
Design and Applications, Research Studies Press, London,
1986.
110. P. H. Ladbrooke, MMIC Design GaAs FETs and HEMTs,
Artech House, London, 1989.
111. F. Grover, Inductance Calculations, Van Nostrand, Prince-
ton, NJ, 1946, reprinted by Dover, New York, 1962.
112. F. J. Schmuckle, The method of lines for the analysis of rect-
angular spiral inductors. IEEE Trans. Microwave Theory
Tech. MTT-41:11831186 (1993).
113. H. M. Greenhouse, Design of planar rectangular microelec-
tronic inductors, IEEE Trans. Parts Hybrids Pack. PHP-
10:101109 (1974).
114. F. H. Terman, Radio Engineer Handbook, McGraw-Hill,
New York, 1943.
115. M. Caulton and H. Sobol, Microwave integrated-circuit tech-
nologya survey, IEEE J. Solid-State Circ. SC-5:292303
(1970).
116. G. D. Alley, Interdigital capacitors and their application
to lumped-element microwave integrated circuits,
(1970).
117. C. S. Aitchison et al., Lumped-circuit elements at microwave
frequencies, IEEE Trans. Microwave Theory Tech. MTT-
19:928937 (1971).
118. L. Young and H. Sobol, eds., Advance in Microwaves, Vol. 8,
Academic Press, New York, 1974, pp. 158159.
119. S. A. Jamison et al., Inductively coupled push-pull amplier
for low cost monolithics microwave ICs, IEEE GaAs IC
Symp. Digest, 1982, pp. 9193.
120. D. Ferguson et al., Transformer coupled high-density circuit
technique for MMIC, IEEE Microwave Millimeter-Wave
Monolithic Circuits Symp. Digest, 1984, pp. 3436.
121. L. Wiemer et al., Computer simulation and experimental
investigation of square spiral transformers for MMIC appli-
cations, IEE Colloq. Comput.-Aided Des. Microwave Cir-
cuits, Digest 99, 1985, pp. V1-5.
122. W. Menzel et al., Compact multilayer lter structures for
coplanar MMICs, IEEE Microwave Guide Wave Lett. 2:497
498 (1992).
123. W. Schwab and W. Menzel, On the design of planar micro-
wave components using multilayer structures, IEEE Trans.
124. M. Engles and R. H. Jansen, Design of quasi-ideal couplers
using multilayer MMIC technology, IEEE MTT-S Int. Mi-
crowave Symp., 1996, pp. 11811184.
125. C. Person et al., Wideband 3 dB/901 coupler in multilayer
thick-lm technology, Electron. Lett. 31:812813 (1995).
126. F. Mernyei, I. Aoki, and H. Matsuura, A novel MMIC cou-
plermeasured and simulated data, IEEE MTT-S Int. Mi-
crowave Symp., 1994, pp. 229232.
127. S. Banba and H. Ogawa, Multilayer MMIC directional cou-
plers using thin dielectric layers, IEEE Trans. Microwave
Theory Tech. MTT-43:12701275 (1995).
128. P. J. Zabinski et al., Example of a mixed-signal global posi-
tioning system (GPS) receiver using MCM-L packaging,
IEEE Trans. Compon. Pack. Manuf. Technol. Part B
CPMT-18:1317 (1995).
129. R. G. Arnold and D. J. Pedder, Microwave characterization of
microstrip lines and spiral inductors in MCM-D technology,
IEEE Trans. Compon. Hybrids Manuf. Technol. CHMT-
15:10381045 (1992).
130. M. Engels and R. H. Jansen, Modeling and design of novel
passive MMIC components with three and more conductor
levels, IEEE MTT-S Int Microwave Symp., San Diego, CA,
1994, pp. 12931296.
131. T. Tokumitsu et al., Three-dimensional MMIC technology for
multifunction integration and its possible application to
masterslice MMIC, IEEE Microwave Millimeter-wave Mono-
lithic Circuits Symp., 1996, pp. 8588.
132. R. G. Arnold and D. J. Pedder, Microwave components in
multichip module (MCM-D) technology, Proc., Microwave
RF, Wembly, London, 1994, pp. 195199.
133. C. Cho and K. C. Gupta, Design methodology for multilayer
coupled line lters, IEEE MTT-S Int. Microwave Symp.,
Denver, CO, 1997, pp. 785788.
134. A. M. Pavio and A. Kikel, A monolithic or hybrid broadband
compensated balun, IEEE MTT-S Int. Microwave Symp.,
1990, pp. 483486.
135. M. Davidovitz, A compact planar magic-T junction with ap-
erture-coupled difference port, IEEE Microwave Guided
Wave Lett. 7:217218 (1997).
136. C. M. Tsai and K. C. Gupta, A generalized model for coupled
lines and its applications to two-layer planar circuits,
(1992).
137. C. M. Tsai and K. C. Gupta, CAD procedures for planar re-
entrant type couplers and three-line baluns, IEEE MTT-S
Int. Microwave Symp., 1993, pp. 10131016.
138. M. Tran and C. Nguyen, Modied broadside-coupled micro-
strip lines suitable for MIC and MMIC applications and a
new class of broadside-coupled band-pass lters, IEEE
Trans. Microwave Theory Tech. 41:13361342 (1993).
139. C. Cho and K. C. Gupta, A new design procedure for single-
layer and two-layer 3-line baluns, IEEE MTT-S Int. Micro-
wave Symp., Baltimore, MD, 1998, pp. 777780.
140. C.-Y. Lee and T. Itoh, Full-wave analysis and design of a new
double-sided branch-line coupler and its complementary
structure, IEEE Trans. Microwave Theory Tech. MTT-
43:18951901 (1995).
141. M. Engels et al., Design methodology, measurement and ap-
plication of MMIC transmission line transformers, IEEE
MTT-S Int. Symp., Orlando, FL, 1995, pp. 16351638.
142. N. Herscovici and D. M. Pozar, Full-wave analysis of aper-
ture-coupled microstrip lines, IEEE Trans. Microwave The-
ory Tech. MTT-39:11081114 (1991).
143. P. R. Haddad and D. M. Pozar, Analysis of two aperture-
coupled cavity-backed antennas, IEEE Trans. Anten. Prop-
ag. AP-45:17171726 (1997).
144. K. C. Gupta, R. Garg, and R. Chadha, CAD of Microwave
Circuits, Artech House, Dedham, MA, 1981.
145. J. A. Dobrowolski, Introduction to Computer Methods for
Microwave Circuit Analysis and Design, Artech House, Nor-
wood, MA, 1991.
146. J. A. Dobrowolski, Computer-Aided Analysis, Modeling, and
Design of Microwave Networks: The Wave Approach, Artech
House, Norwood, MA, 1996.
147. M. Hirano, Three-dimensional passive circuit technology for
ultra compact MMICs, IEEE Trans. Microwave Theory Tech.
MTT-43:28452849 (1995).
148. I. Toyoda et al., Three-dimensional masterslice MMIC on Si
substrate, IEEE Trans. Microwave Theory Tech. MTT-
45:25242530 (1997).
149. J. A. Navarro and K. Chang, Integrated Active Antennas and
Spatial Power Combining, Wiley, New York, 1996.
150. R. A. York and Z. B. Popovic, Active and Quasi-Optical Ar-
rays for Solid State Power Combining, Wiley, New York,
1997.
151. P. M. Watson and K. C. Gupta, Design and optimization of
CPW circuits using EM-ANN model for CPW components,
(1997).
MICROWAVE DETECTORS
RONALD E. HAM
Consultant
Microwave energy cannot be efciently detected with
equipment used for lower-frequency applications. Micro-
wave-frequency voltage changes are too rapid to be cap-
tured on any but the fastest sampling oscilloscopes, and
response time and parasitics of common test meters rend-
er them useless. Microwave detectors generally measure
power in the controlled impedance environments afforded
by transmission lines and waveguides. The system imped-
ance is known, so voltage and current magnitudes can be
derived. With the judicious application of other microwave
components such as power splitters and quadrature cou-
plers, phase information can be extracted from multiple
power measurements; therefore, full characterization of a
system or component is possible with only the microwave
detector to convert microwave signal characteristics to the
readily observable and measurable quantities of low-fre-
quency voltages and currents.
Microwave detectors sense the amplitude or power of a
signal, or the change in the amplitude or power of a signal.
They are used for the determination of signal power level
and for recovery of information placed upon the micro-
wave signal amplitude. The microwave detector circuit
(Fig. 1) consists of a matching circuit to match the imped-
ance of the circuit being measured to the impedance of the
detector itself, along with an output signal conditioning
circuit to convert the detector output to a voltage or cur-
rent that is proportional to the microwave signal input.
Ideally, the detector impedance directly matches the sys-
tem impedance, and the matching circuit is not required.
Microwave detectors can be separated into two catego-
ries: thermally dependent detectors and diode detectors.
Thermally dependent detectors absorb the incident micro-
wave signal in a detector element, and the microwave en-
ergy is converted into heat energy. The increase in
temperature of the detector element is directly detected,
or a change in a physical parameter of the detector ele-
ment due to the change in temperature, normally resis-
tance, is detected. The diode detector recties the
microwave signal, clipping either the positive or negative
voltage of the alternating voltage waveform, and averages
the resultant monopolar signal in a capacitive circuit to
produce a DC signal with an amplitude related to the
microwave signal amplitude.
The response time of the thermally dependent detector
is a function of the thermal mass being heated by the in-
cident energy and the thermal resistance of the detector to
its surrounding environment. In general, the response
time of a thermally dependent detector is slow compared
to the response time of a diode detector and is on the order
of milliseconds to seconds. Thermally dependent detectors
are used primarily for the measurement of the power level
of steady-state signals or to measure the average value of
the power of a time-varying signal such as a pulsed radar
transmitter output.
The response time of a diode detector can be within a
few periods of the microwave signal being detected; hence,
diode detectors can recover information placed in ampli-
tude variations on a signal in addition to measuring the
power of steady-state signals.
Small-signal thermally dependent detectors with a
bandwidth of less than 100Hz typically operate over a
power range of 10 dBm to about 50 dBm, while diodes
limited to the same bandwidth are useful from about
10 dBm to less than 80 dBm. Insertion of attenuators or
couplers between the power source and the detector
facilitates measuring much higher powers; however, mea-
surement of very small amounts of power generally ne-
cessitates insertion of ampliers or frequency conversion
to lower frequencies where more sensitive techniques can
be used to detect the signal.
1. THERMALLY DEPENDENT DETECTORS
Thermally dependent detectors depend on the heating of
the detector element by the incident microwave energy.
The heating is proportional to the electrical energy dissi-
pated in the detector element. Bolometers and thermocou-
ples are typical of thermally dependent microwave
detectors. Because absorption of the available power in
the incident signal depends on a proper impedance match
between the transmission media and the absorbing detec-
tor circuit, the detector must have a well-established and
controlled impedance.
Microwave
signal
input
Impedance
matching
circuit
Microwave
detector
Signal
conditioning
Voltage
or
current
output
Figure 1. The impedance of the microwave detector should be matched to the impedance of the
system being measured.
2736 MICROWAVE DETECTORS
Next Page

12e

Încărcat de

Informații document

Descriere originală:

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

12e

Încărcat de

Drepturi de autor:

Formate disponibile

MICROWAVE CIRCUITS

Figure 2. Field patterns of TE

Figure 7. Two other designs of waveguide directional coupler

S-ar putea să vă placă și